Growth strategies in the squid Loligo vulgaris from Portuguese waters

Abstract

The growth of the European squid Loligo vulgaris in northwest Portuguese waters is described and the influences of gender and hatching season analysed, based on statolith readings from individuals of a wide range of sizes. Male and female growth follows different models, males attaining a higher length-at-age than females. Males display increasing growth rates irrespective of the hatching season, but the length-at-age is higher in animals hatched during the warm season. Females may exhibit asymptotic growth or not, depending upon the environmental conditions to which they are exposed through their life cycle. Although growth rates after hatching are lower in females hatched during the cold season, favourable feeding and temperature conditions during the following spring and summer months contribute not only to increase growth rates but also to delay sexual maturation. The higher length-at-age of squid hatching in the warm season, observed in both genders, provides evidence that the temperature close to hatching has a significant impact on the size of juveniles and subadults. However, there is also strong evidence that throughout their life, environmental conditions continue to play an important role in growth rates and in defining the shape of growth.

Full text

Downloaded By: [University of Aberdeen] At: 15:42 2 March 2007
ORIGINAL ARTICLE
Growth strategies in the squid Loligo vulgaris from Portuguese waters
ANA MORENO
1
, MANUELA AZEVEDO
1
,JOA
˜O PEREIRA
1
& GRAHAM J. PIERCE
2
1
Instituto Nacional de Investigac
¸a
˜oAgra
´ria e das Pescas, INIAP-IPIMAR, Av. Brası
´lia, 1400-038 Lisbon, Portugal;
2
School of Biological Sciences (Zoology), College of Life Sciences and Medicine, University of Aberdeen, Tillydrone Avenue,
Aberdeen AB242TZ, UK
Abstract
The growth of the European squid Loligo vulgaris in northwest Portuguese waters is described and the influences of gender
and hatching season analysed, based on statolith readings from individuals of a wide range of sizes. Male and female growth
follows different models, males attaining a higher length-at-age than females. Males display increasing growth rates
irrespective of the hatching season, but the length-at-age is higher in animals hatched during the warm season. Females may
exhibit asymptotic growth or not, depending upon the environmental conditions to which they are exposed through their life
cycle. Although growth rates after hatching are lower in females hatched during the cold season, favourable feeding and
temperature conditions during the following spring and summer months contribute not only to increase growth rates but
also to delay sexual maturation. The higher length-at-age of squid hatching in the warm season, observed in both genders,
provides evidence that the temperature close to hatching has a significant impact on the size of juveniles and subadults.
However, there is also strong evidence that throughout their life, environmental conditions continue to play an important
role in growth rates and in defining the shape of growth.
Key words: Age, environmental factors, growth models, Loligo vulgaris, squid
Introduction
Cephalopod growth is highly variable, similar to
other aspects of their biology. This variability is
believed to be a combination of intrinsic variability
and the influence of several environmental factors, of
which temperature and food intake have major roles
(reviewed in Forsythe & Van Heukelem 1987;
Jackson 1994).
Several studies investigating the impact of tem-
perature on variability in size-at-age were conducted
in the last decade in wild populations (e.g. Jackson et
al. 1997; Hatfield 2000; Jackson & Moltschaniwskyj
2002) to corroborate the Forsythe hypothesis (For-
sythe 1993) or Forsythe effect (Forsythe 2004),
which states that because hatching in many species
occurs over a period of continually changing tem-
peratures, each cohort of hatchlings encountering
warmer temperatures will grow significantly faster
than those that hatched only weeks previously. Other
studies in wild squid populations revealed significant
differences in size-at-age between geographical areas
(Jackson & Moltschaniwskyj 2002; Chen & Chiu
2003) and between seasonal cohorts (Arkhipkin
et al. 2000; Villegas 2001; Pecl 2004). These
differences in growth were mainly attributed to the
geographical or seasonal prevailing temperatures
during the early life stages.
Loligo vulgaris is a near-shore species distributed
along Northeastern Atlantic waters and the Medi-
terranean, where it is commercially exploited. It is
widely distributed along the Portuguese continental
shelf, displaying yearlong spawning and recruitment
(Moreno et al. 2002). On the northwest Portuguese
coast, L. vulgaris occurs mainly from the shore to the
100 m isobath. In contrast to most other loliginid
populations, spawners and juveniles co-occur, both
in time and space, and are available to fisheries all
year round (Cunha et al. 1995; Moreno 1998).
Nevertheless, some horizontal short-range mixing of
animals from different spawning grounds may occur,
as there is no significant evidence of population
Correspondence: A. Moreno, Instituto Nacional de Investigac
¸a
˜oAgra´ria e das Pescas, INIAP-IPIMAR, Av. Brası´lia, 1400-038 Lisbon,
Portugal. E-mail: amoreno@ipimar.pt
Published in collaboration with the University of Bergen and the Institute of Marine Research, Norway, and the Marine Biological Laboratory,
University of Copenhagen, Denmark
Marine Biology Research, 2007; 3: 49 59
(Accepted 13 November 2006; Printed 5 March 2007)
ISSN 1745-1000 print/ISSN 1745-1019 online #2007 Taylor & Francis
DOI: 10.1080/17451000601129115

Downloaded By: [University of Aberdeen] At: 15:42 2 March 2007
substructuring (Pierce et al. 1994b; Anonymous
2004).
Currently, no formal assessment of this resource is
undertaken. Some exercises with cohort analysis
(Royer et al. 2002; Challier et al. 2006a) and
depletion methods (Young et al. 2004) have been
conducted for loliginid squids, but for the most part,
conventional models used in the assessment of fin
fish stocks are not applicable (Pierce & Guerra
1994). Nevertheless, if stock size predictions are to
partly rely on temperature data (Robin & Denis
1999; Pierce & Boyle 2003; Chen et al. 2006), the
understanding of age structure, growth, and its
variability under fluctuating environmental condi-
tions is of fundamental relevance.
There is a general consensus among squid re-
searchers (Jackson 1994) that squid growth does
not follow the generalized von Bertalanffy model
(Bertalanffy 1938) often adopted to describe fish
growth. Cephalopod growth is generally described
as continuous and non-asymptotic (Jackson &
Moltschaniwskyj 2002). Nonetheless, the growth of
many ommastrephid squid species has been mod-
elled using Gompertz or logistic models (Todarodes
sagitattus , Arkhipkin et al. 1999; Todaropsis eblanae ,
Arkhipkin & Laptikhovsky 2000; Illex coindetii ,
Arkhipkin et al. 2000; Illex illecebrosus , Hendrickson
2004; Dosidicus gigas , Markaida et al. 2004), which
suggests an inflexion in growth at some point. It
seems that the power model best fits most Loligo
size-at-age data (e.g. L. vulgaris , Arkhipkin 1995; L.
pealeii, Macy & Brodziak 2001; L. gahi , Villegas
2001; L. opalescens , Jackson & Domeier 2003).
Other models, however, best fitted the age data of
some of the above species when applied to popula-
tions of different geographical areas, including
exponential models (L. vulgaris , Rocha & Guerra
1999; L. forbesi, Challier et al. 2006b) or double
exponential models (L. vulgaris , Natsukari &
Komine 1992).
Whether squid grow asymptotically or not was not
an issue in the present study. Instead, our aim was to
analyse the variability in the type of growth model
within the population, between genders and seasonal
cohorts and, in particular, whether there was an
inflexion in growth and why. An analysis of age and
size at maturity in L. vulgaris from northwest
Portuguese waters has already revealed that varia-
bility is closely related to the environmental condi-
tions experienced by individuals through their life
cycle, namely those affecting the maturation rates of
females (Moreno et al. 2005). The aim of this study
was to further the understanding of life cycle
plasticity in L. vulgaris by analysing their growth
variability.
Material and methods
Samples were collected monthly, between January
1993 and December 1994, along the northwest
Portuguese coast (38830’ to 428N). Every month,
samples were obtained from either the trawl or the
seine fisheries both operating within the same
geographical area. In order to increase the sample
size of large squid and to include pre-recruits in the
sizeage range, additional samples from the jig
fisheries and survey cruises carried out in the same
area between March 1993 and March 2000 were
also used. Dorsal mantle length (ML, mm), gender,
age and maturity stage were determined in 435
individuals (23 unknown gender, 193 females and
219 males). Maturity stages 1, 2 3 and 4 5 were
used to classify individuals as immature, maturing
and mature, respectively, following the five-stage
maturity scale described in Boyle & Ngoile (1993).
The individuals of unknown gender were all small,
immature animals with insufficient gonad develop-
ment to allow the gender to be determined. Age was
determined by increment counting in the statoliths
following the methodology detailed in Moreno
(2002). Deposition of increments in the statoliths
was assumed to be daily, as was validated for
paralarvae of this species (Villanueva 2000). Ageing
precision was evaluated by independent readings
(same reader, distinct reading date) in 50 statoliths
[details in Moreno et al. (2005)], showing high
precision (coefficient of variation 
/3.2%). The
hatching date was calculated as the capture date
minus the age. For some of the analyses, ages were
grouped into 30-day classes. To examine the effect
of gender on growth, the length-at-age data of the
unknown gender specimens were added to data for
each gender to increase the length age range,
assuming no significant gender-related differences
in smaller specimens.
The average sea surface temperature (SST) in the
first 3 months of the life of each specimen was
computed. To explore the effect of SST during early
life on growth, while controlling for hatch month
and year, we fitted a Gaussian generalized additive
model (GAM) to (sqrt transformed in the case of
males) length-at-age data. SST during the first 3
months of life and month were treated as continuous
variables and the year (1992, 1993, 1994, other) as a
nominal variable. Degrees of freedom for smoothers
were estimated using cross-validation. This preli-
minary analysis showed a significant effect of tem-
perature on length-at-age (P 
/0.001 for both
genders). To maximize SST differences, two groups
of squid were selected, named the cold cohort (CC,
average SSTB
/158C) and the warm cohort (WC,
average SST
/16.58C), regardless of the hatching
50 A. Moreno et al.

Downloaded By: [University of Aberdeen] At: 15:42 2 March 2007
year or season. Weekly SST data for northwest
Portuguese waters (38.5 41.58N and 8.5 9.58W)
were extracted from the integrated Global Ocean
Services System Meteorological Center ‘‘IGOSS
nmc’’ database (http://ingrid.ldgo.columbia.edu/
SOURCES/.IGOSS/.nmc/. monthly/.sst/; Reynolds
& Smith 1994) on a 18latitude 
/18longitude grid.
Analysis of covariance (ANCOVA) was used to test
the effect of the cohort on the size age relationship of
each gender, using age as the covariate. Because of
the unequal sample size by age class between cohorts,
the average length-at-age of the CC and WC groups
was compared by gender using a data set generated
by random subsampling of an equal number of
specimens by age class in each cohort.
Six models were fitted to the (individual) length-
at-age data, namely linear, exponential, power,
Gompertz, logistic and von Bertalanffy. The para-
meters of the models were estimated by the least
squares method (linear and intrinsic linear models)
and the iterative simplex method (non-linear mod-
els). The best-fit model was determined by the
highest goodness of fit, the examination of residuals
for any systematic pattern, and the reliability of the
parameter estimates. Growth model fitting and
ANCOVA were performed with R using the stats
package (R Development Core Team 2006). GAMs
were fitted using the R interface statistical package
BRODGAR.
Daily growth rates (DGR), by gender and hatch-
ing cohort, and instantaneous growth rates (G) for
each gender were determined in 30-day time inter-
vals, after Ricker (1979):
DGR(MLi1MLi)=(ti1ti);
G(ln(MLi1)ln(MLi))=(ti1ti)100
where ML
i
and ML
i1
and t
i
and t
i1
are the ML
(mm) and age (days) at the beginning and the end
of each age class t, respectively. For time intervals
t
/30 days, the arithmetic means of DGR and G
were calculated (avgDGR or avgG).
Results
Age composition
The ML of aged individuals ranged from 21 to 546
mm (21 to 72 mm for unknown gender, 31 to 332
mm for females and 64 to 546 mm for males). The
youngest squid was 87 days old. Gender could be
determined in specimens from 120 days old. The age
of the CC squid ranged from 93 to 387 days (93 to
109 days for unknown gender, 181 to 357 days for
females and 165 to 387 days for males) and the age
of the WC squid ranged from 88 to 446 days (88 to
213 days for unknown gender, 140 to 439 days for
females and 154 to 446 days for males). In general,
the maximum age observed was 439 days in females
and 446 days in males and thus the maximum
longevity was estimated to be ca . 15 months in
both genders, although the majority of specimens
were less than 1 year old.
Preliminary GAM modelling
The GAMs fitted to length-at-age data demon-
strated that SST in the first 3 months of life
significantly affected growth in both males and
females (PB
/0.001 in both cases). In both genders,
the shape of the smoother for the effect of tempera-
ture indicated that squid reached higher length-at-
age when SST was higher, at least up to around 178C
(Figure 1). There were also significant effects of
hatch month in both genders (P B
/0.001 in females,
P
/0.001 in males), with the lowest length-at-age
(for a given temperature experienced in early life)
achieved by squid hatching during June to Septem-
ber. (Because these months tend to be warm
months, squid hatched in these months would still
on average achieve larger size-at-age.) Females
hatched in 1993 reached larger lengths-at-age than
females hatched in 1992 (P 
/0.002), but there were
no other significant year effects. For the subsequent
analysis, hatch month and year effects on growth
were ignored.
It should be noted that, as in most growth studies,
we infer patterns of individual growth from single
measurements of length-at-age in a large number of
individuals. Obviously, the more individual growth
trajectories vary, the greater chance there is that the
composite growth model is a poor representation of
the growth in any one individual. This caveat should
be borne in mind in any discussion of alternative
growth models.
Gender differences in growth
ML was significantly correlated with age (P 
/0.001).
This correlation increased when computed by gen-
der, showing that gender is a source of variability in
length-at-age. The effect of gender on growth was
significant (ANCOVA, F
1,409
/7.36, P/0.007),
showing that males grew faster than females. The
models’ goodness of fit was similar in females (0.75 
0.82) and in males (0.74 0.78). The scatter-plot for
females (Figure 2A) showed an inflexion, suggesting
that an asymptotic model could have described
growth better (Table I), and this was corroborated
by the systematic patterns detected in the residuals of
the non-asymptotic models, namely concerning
smaller and/or larger animals. Although the goodness
of fit for both Gompertz and logistic models was of
the same magnitude (0.82), the logistic model was
Loligo vulgaris growth strategies 51

Downloaded By: [University of Aberdeen] At: 15:42 2 March 2007
adopted as better describing female growth, given
that the estimated asymptotic ML was closer to the
maximum observed ML (logistic: ML

/335 mm,
standard error 
/23.3 mm; Gompertz: ML

/
415 mm, standard error /55.9 mm) and the stan-
dard error of the estimated parameters was lower.
The ML-at-age data for males did not indicate any
growth inflexion, thus asymptotic models presented
non-reliable parameter estimates. The power model
was adopted (Figure 2B, Table I) as the best fitting
growth model, with higher goodness of fit (0.78) and
no systematic pattern in the residuals, in relation to
the linear and exponential models.
Female DGR increased from 0.42 mm day
1
in the fourth month of life (90 120 days) to a
maximum of 1.16 mm day
1
in the ninth month
(240270 days), around the growth inflexion
point (257 days), and subsequently decreasing
(Figure 3). On the other hand, males showed a
steady increase in DGR until death. DGR
were similar between genders until squid were
240 days old (AvgDGR
90 240F
/0.77 mm day
1
and AvgDGR
90 240M
/0.83 mm day
1
). Gender
differences became more apparent when the squid
were 270 days old (ca. 5 mm difference in ML),
reaching a difference of ca. 65 mm in ML when they
were about 1 year old and a difference of ca. 150 mm
ML when squid reached 420 days old. In this period
of life, the growth rate of males was much higher
(AvgDGR
240 450M
/1.70 mm day
1
) than that
of females (AvgDGR
240 450F
/0.77 mm day
1
).
Between 90 and 450 days of age, females grew at
an average rate (avgDGR) of 0.77 mm day
1
and
males at an average rate of 1.34 mm day
1
. The
instantaneous growth rate (G) decreased until death
Figure 1. Generalized additive model of length-at-age in Loligo vulgaris . Partial effects (solid line) and 95% confidence limits (broken lines)
of smooth terms of SST in the first 3 months of life and hatch month for females (upper panel) and males (lower panel). SST, sea surface
temperature.
0
100
200
300
400
500
600
0 50 100 150 200 250 300 350 400 450 500
Age (days)
ML (mm)
immature
maturing
mature
model
Females
0
100
200
300
400
500
600
0100 200 300 400 500
Age (days)
ML (mm)
Males
(a)
(b)
Figure 2. The relationship between age and mantle length for
Loligo vulgaris females (A) and males (B). The fitted curves are the
logistic model for females and the power model for males.
52 A. Moreno et al.

Downloaded By: [University of Aberdeen] At: 15:42 2 March 2007
in both genders. This growth rate was always higher
in males (AvgG
90 450M
/0.86% day
1
) than in
females (AvgG
90 450M
/0.65% day
1
), explaining
the larger sizes reached by males.
Cohort differences in growth
Growth variability in the CC was higher than in the
WC for both females and males. ANCOVA tests
confirmed that the effect of hatching season on
growth was significant for both genders (females,
F
1,83
/17.1, P/0.001; males, F
1,111
/8.1, PB/
0.01), showing that the mean length-at-age, in the
period between 180 days of age and the end of the life
cycle, was higher in the WC (175 mm for females,
n
/43 and 224 mm for males, n /59) than in the CC
(145 mm for females, n 
/43 and 174 mm for males,
n
/55). In squid younger than 180 days, no sig-
nificant differences in mean ML were found between
cohorts (WC: 56 mm, n 
/30; CC: 42 mm, n /11)
(ANCOVA F
1,38
/2.5, P/0.05).
The best fitted growth models for each seasonal
female cohort showed differences between cohorts
(Table II). The growth of WC females was asymp-
totic, following a logistic model with an inflexion
point at 237 days and asymptotic ML at 303 mm,
whereas the growth of CC females followed a power
model (Figure 4A, B). The growth models indicated
that after 350 days, CC females are larger. However,
all observed CC females were less than 1 year old.
The fit of asymptotic models to CC females proved
inadequate, yielding very high ML

values and
growth inflexions close to the maximum age. On
the other hand, in WC females, growth was clearly
asymptotic (even if the three oldest specimens
were discarded) and the non-asymptotic models
showed clear systematic patterns of the residuals
and lower goodness of fit (Figure 5A, B). The
logistic model was adopted because it showed higher
r
2
and a slightly better fit in the lower ages than the
Gompertz model.
In the case of males, there was no evidence of
asymptotic growth in either cohort and non-
Table I. Parameter estimation of the general growth models for females and males.
Female growth model ML

kt
i
Estimate 335.4 0.014 256.5
Logistic model Standard error 21.7 0.001 10.2
ML/ML

/(1/exp(/k(t/t
i
))) t-value (df /213) 15.5 11.8 25.0
r
2
/0.82 P level 0.00 0.00 0.00
Lower confidence limit 292.6 0.012 236.3
Upper confidence limit 378.1 0.017 276.7
Male growth model a b
Estimate 0.003 1.98
Power model Standard error 0.001 0.07
ML
/a.t
b
t-value (df/240) 2.4 27.3
r
2
/0.78 P level 0.02 0.00
Lower confidence limit 0.001 1.84
Upper confidence limit 0.005 2.12
0
0.5
1
1.5
2
Age class (days)
DGR (mm/d)
0
0.5
1
1.5
2
G (% mm/d)
F (DGR) M (DGR) F (G) M (G)
90 120 180150 270210 360330300 450420390240
Figure 3. Variation in daily growth rate (DGR) and instantaneous growth rate (G) through the life cycle for Loligo vulgaris females (F) and
males (M). Growth rates were calculated in 30-day time intervals.
Loligo vulgaris growth strategies 53

Downloaded By: [University of Aberdeen] At: 15:42 2 March 2007
asymptotic models were adopted to describe the
growth of both seasonal cohorts (Table II). The
growth of CC males (Figure 6A) was better fitted by
an exponential model (acceptable residual pattern
and higher goodness of fit), whereas the growth of
WC males (Figure 6B) was better fitted by a power
model (all other models present systematic patterns
in residuals and a lower goodness of fit). CC males
had lower ML-at-age than WC males until 360 days
old, and higher ML-at-age above this age.
Monthly variation in growth rates was such that
the DGR of WC females increased sharply until a
maximum was reached during the eighth month
(corresponding to the growth inflexion point) and
decreasing thereafter (Figure 7A). On the other
hand, the DGR of CC females increased steadily
until the end of the life cycle. The WC growth
rate was higher until squid were 270 days
old (AvgDGR
90:270WC
/0.96 mm day
1
,Avg
DGR
90:270CC
/0.74 mm day
1
) and the opposite
was true after that age (AvgDGR
270:360WC
/
0.84 mm day
1
, AvgDGR
270:360CC
/1.13 mm
day
1
). As in females, the DGR of WC males
(Figure 7B) was higher until they reached 270
days old (AvgDGR
90:270WC
/0.91 mm day
1
,
AvgDGR
90:270CC
/0.66 mm day
1
) and the oppo-
site was verified after that age (AvgDGR
270:360WC
/
1.47 mm day
1
, AvgDGR
270:360CC
/1.88 mm
day
1
).
Discussion
The preliminary GAM modelling indicated that, for
squid that experienced SST up to around 16.58C
during the first 3 months of life, the effect of SST on
length-at-age was positive. Above this temperature,
there was no further consistent increase in length-at-
age related to increased SST. The analysis also
revealed significant effects of hatching month and
year, even after the effect of SST was taken into
account. For the purpose of fitting parametric
growth models, we ignored these latter effects and
defined two ‘‘cohorts’’ representing cold (CC) and
warm (WC) portions of the SST range.
In spite of the similar sizes reached in both
cohorts, a longer life span was detected in the WC
squid, which spend their adult life in colder waters
than those occupied by the CC squid. This leads to
the hypothesis of intraspecific geographical variation
in life span, with a tendency for lower longevity in
the populations living in warmer waters. This is
corroborated by the higher maximum age of L.
vulgaris in the northwest Portuguese waters (ca .15
months) than that observed in the northwest African
Table II. Parameter estimation of the adopted growth models for the females of the cold cohort (FCC), females of the warm cohort (FWC),
males of the cold cohort (MCC) and males of the warm cohort (MWC).
FCC growth model a b
Estimate 0.009 1.75
Power model Standard error 0.37* 0.07
ML
/a.t
b
t-value (df/76) /12.6 25.7
r
2
/0.90 P level 0.00 0.00
Lower confidence limit 0.004 1.61
Upper confidence limit 0.019 1.88
FWC growth model ML

kt
i
Estimate 302.6 0.019 236.8
Logistic model Standard error 18.7 0.002 8.1
ML
/ML

/(1/exp(/k(t/t
i
))) t-value (df/82) 16.2 10.8 29.4
r
2
/0.87 P level 0.00 0.00 0.00
Lower confidence limit 265.3 0.015 220.8
Upper confidence limit 339.9 0.022 252.9
MCC growth model a b
Estimate 17.5 0.008
Exponential model Standard error 3.1 0.001
ML
/a.exp(b.t) t-value (df/78) 5.7 14.3
r
2
/0.75 P level 0.00 0.00
Lower confidence limit 11.4 0.007
Upper confidence limit 23.7 0.009
MWC growth model a b
Estimate 0.007 1.84
Power model Standard error 0.003 0.09
ML
/a.t
b
t-value (df/97) 2.0 21.5
r
2
/0.85 P level 0.045 0.00
Lower confidence limit 0.000 1.67
Upper confidence limit 0.013 2.00
*Standard error of ln(a), parameter estimation with ln transformed data.
54 A. Moreno et al.

Downloaded By: [University of Aberdeen] At: 15:42 2 March 2007
coast (1013 months; Arkhipkin 1995; Raya et al.
1999), where the average annual SST is ca .48C
higher (Moreno et al. 2002).
Early growth
The analysis of growth rates (DGR) showed that
juvenile squid grow to 75 mm ML in 170 days. This
growth rate is slower than that observed for the same
species by Turk et al. (1986) in captivity (75 mm in
112140 days), but is similar to what Arkhipkin
(1995) observed in nature on the northwest African
coast (7080 mm in 150 160 days). Higher growth
rates up to 270 days of age were detected in squid
hatched under warm temperatures, reflecting the
influence of environmental factors. Water tempera-
ture itself may be the most important factor because
of accelerated metabolism. Other factors, such as
food availability, may be limiting, whereas others,
such as the photoperiod, may act as cues to trigger
internal responses. Zooplankton abundance on the
northwest Portuguese coast is, in general, higher
between May and October and lower between
December and February (Cunha 1993), thus pro-
viding more favourable feeding conditions for the
early stages of the WC than for those of the CC. In
culture experiments, Villanueva (2000) observed a
higher ML (double after 50 days) in L. vulgaris
0
50
100
150
200
250
300
350
0 100 200 300 400 500
Age (days)
ML (mm)
immature
maturing
mature
model
0
50
100
150
200
250
300
350
0 100 200 300 400 500
Age (days)
ML (mm)
immature
maturing
mature
model
(a)
(b)
Figure 4. The relationship between age and mantle length for
Loligo vulgaris females of (A) the cold cohort (CC) and (B) the
warm cohort (WC). The fitted curves are the power model for the
CC and the logistic model for the WC.
–100
0
100
0 100 200 300 400
Predicted values
Residual values
Logistic model
–100
0
100
0 100 200 300 400
Predicted values
Residual values
Power model
(b)
(a)
Figure 5. Example of residual versus predicted values from the
fitting of (A) an asymptotic growth model or (B) a non-asymptotic
growth model to data for the Loligo vulgaris warm cohort females.
0
100
200
300
400
500
600
0 100 200 300 400 500
Age (days)
ML (mm)
immature
maturing
mature
model
0
100
200
300
400
500
600
0 100 200 300 400 500
A
g
e (days)
ML (mm)
immature
maturing
mature
model
(b)
(a)
Figure 6. The relationship between age and mantle length for
Loligo vulgaris males of (A) the cold cohort (CC) and (B) the
warm cohort (WC). The fitted curves are the exponential model
for the CC and the power model for the WC.
Loligo vulgaris growth strategies 55

Downloaded By: [University of Aberdeen] At: 15:42 2 March 2007
hatched and maintained under high temperatures
(19.58C) compared with those hatched and main-
tained (also for 50 days) under low temperatures
(12.28C). In our study, in spite of the differences in
growth rates between cohorts, the observed differ-
ences in the size-at-age of young squid were not
statistically significant.
One explanation for this is that the WC squid are
smaller in size-at-hatching, masking the result of a
higher early growth rate at higher temperatures (Pecl
et al. 2004). After some time, larger hatchlings
growing slower may reach similar sizes to those
reached by smaller hatchlings growing faster. More-
over, in a study based on wild caught specimens,
individual variability may also mask the differences
between cohorts, particularly when the absolute
difference in growth is necessarily of a small magni-
tude due to the smaller size of the specimens.
Nevertheless, it should be noted that our ability to
detect significant differences might have been com-
promised by the low sample size of the young squid
from the CC; thus, further research may be required
to support the above hypotheses.
Gender differentiation
The short life span of squid and their ability to
generate new muscle fibres throughout their lives
(Moltschaniwskyj 1994) have been cited as the main
features supporting continuous and non-asymptotic
growth. In fact, the majority of growth models used
so far to describe the growth of Loligo species were
linear or near linear, with no inflexion point or
asymptotic length. Based on a wide range of
individual sizes and ages from our study, L. vulgaris
growth along the northwest Portuguese coast dif-
fered between genders. Female growth clearly
showed an inflexion point followed by a decrease
in the growth rate, best described by a logistic
model, whereas males exhibited growth according
to a power model. Although females grew asympto-
tically (even if they only approached the asymptotic
ML by the end of their life span), both genders
demonstrated continuous growth. Raya et al. (1999)
also observed a decrease in the growth rate of L.
vulgaris females in the northwest African coast
beyond a certain age and, because sampling data
included few old females, the hypothesis of asymp-
totic growth in females was not excluded. More
importantly, however, our study revealed that apart
from gender differences in size-at-age, mainly in the
second half of the life cycle, the growth strategy of
L. vulgaris is also sexually differentiated. Depending
upon the environmental conditions to which they are
exposed through their life cycle, females may exhibit
asymptotic growth or not. On the other hand,
male growth rates always show an increasing trend
and their reaction to environmental conditions is
revealed mainly by the magnitude of that increase. A
description of these strategies is presented below and
the reason why they differ is discussed in light of
endogenous and exogenous factors.
Growth strategies in males versus females
Differences in the ML-at-age between genders at a
certain age were previously reported for L. vulgaris
(e.g. Arkhipkin 1995; Rocha & Guerra 1999). In our
results, gender differences in growth were detected
mainly after the age of 8 months, when males
continued to increase growth rates, whereas some
of the females experienced a growth inflexion. The
inflexion point of female growth (8.6 months)
coincided with their t
50%
(8.7 months; Moreno et
al. 2005), indicating that maturation was closely
related to the decrease in the somatic growth rate.
This is similar to the results obtained by Smith et al.
(2005) for L. forbesi and agrees with the gender
difference in squid reproductive investment that can
be seen from the very high differences between the
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
Age class (days)
DGR (mm/d)
FCC
FWC
90 120 150 180 210 240 270 300 330 360
0
0.5
1
1.5
2
2.5
3
3.5
A
g
e class (days)
DGR (mm/d)
MCC
MWC
90 120 150 180 210 240 270 300 330 360 390
(b)
(a)
Figure 7. Variation in daily growth rate (DGR) through the life
cycle for Loligo vulgaris females (A) and males (B) of the cold
cohort (FCC and MCC) and the warm cohort (FWC and
MWC). Growth rates were calculated for 30-day time intervals.
56 A. Moreno et al.

Downloaded By: [University of Aberdeen] At: 15:42 2 March 2007
gonadosomatic indices of males and females
(Stearns 1992).
In wild populations, a faster growth of summer
hatchers and a slower growth of winter hatchers was
reported by Raya et al. (1999) in L. vulgaris from the
Saharan Bank. Hatfield (2000) and Natsukari et al.
(1988) also observed a significantly higher length-at-
age in squid hatched in the warmer season in relation
to those hatched in the cold season in wild popula-
tions of L. gahi in Patagonia and Photololigo edulis in
Japan, respectively.
When analysing the growth of each cohort, we
observed that environmental conditions throughout
the life cycle may also strongly influence the growth
strategy of either gender, which fits distinct growth
models. The growth inflexion of the general female
population was not detected in the females of the
CC, in which growth after 90 days followed a power
model. As a result of the low early growth rates, the
minimum size at maturity (15 cm; Moreno et al.
2005) is attained later in life in the CC, when these
squid are approximately 8 months old, during the
summer months. The summer environmental con-
ditions of higher temperature, food availability and
day-length favour the delay in maturation and high
growth rates. Because significant reproductive in-
vestment is only made later in life, after 10 months of
age (Moreno et al. 2005), and the transfer of
metabolic resources from somatic to reproductive
growth can be compensated by the favourable
autumn conditions when many prey items become
more abundant (IPIMAR unpubl. data; Pierce et al.
1994a), there is no significant change in growth
pattern. The females of the WC, on the contrary,
initially grow under favourable spring and summer
environmental conditions, maturing younger, as
soon as they reach the minimum size threshold,
which explains the lack of overlap between maturing
and mature individuals shown in Figure 4B. How-
ever, as this minimum size at maturity happens
mainly during winter and early spring, under more
adverse environmental conditions, they then experi-
ence a shortage of resources when they have to spend
more energy in maturation. The association of both
exogenous and endogenous factors thus contributes
to the observed decrease in their growth rates
beyond 7 months of age, i.e. from the age at maturity
when reproductive investment increases significantly
(t
50%
WC/7.3 months; Moreno et al. 2005).
Ultimately, the exogenous factors appear to be the
most important factors affecting the growth of
juvenile (immature) females. The growth of subadult
and adult (maturing and mature) females is more
dependent on a coupling between exogenous and
endogenous factors, which regulate the maturation
process. Male growth rates in both seasonal cohorts
showed an increasing trend. Male somatic growth is
little affected by the maturation process and, be-
cause of this, males are able to adjust resource
allocation to suit their ambient environment more
closely than females (Steer & Jackson 2004). Thus,
the magnitude of their growth rate is more directly
related to the influence of environmental conditions.
The ‘‘second phase of growth’’ of cephalopods
has been frequently considered as a temperature-
independent phase (Grist & des Clers 1999). How-
ever, from our results, it seems quite evident that
environmental conditions, although not as dramati-
cally as in the paralarval phase, play a significant
role in the rate and shape of growth of squid through
the life cycle. The existence of differences in
growth rates, rates of maturation and longevity,
among others, mediated by environmental factors,
emphasizes the need to take environmental data
into consideration in fisheries models, particularly
in times of acknowledged environmental global
change.
Acknowledgements
The authors thank Pedro Mendonc
¸a for assistance
with the biological sampling and Cristina Castro
who helped in statolith processing. Part of this
study was carried out under the European Commis-
sion-funded projects Eurosquid II (AIR-CT92-
0573) and CEPHVAR (FAIR-CT96-1520). Some
of the samples were obtained under the PNAB/Data
Collection programme. We also thank the anon-
ymous referees for their helpful comments on the
draft manuscript.
References
Anonymous 2004. Report of the Working Group on Cephalopod
Fisheries and Life History (WGCEPH). ICES CM 2004/G:02,
Ref. ACFM, ACE.
Arkhipkin AI. 1995. Age, growth and maturation of the European
squid Loligo vulgaris (Myopsidae, Loliginidae) on the west
Saharan Shelf. Journal of the Marine Biological Association of
the UK 75:593 604.
Arkhipkin AI, Jereb P, Ragonese S. 2000. Growth and maturation
in two successive seasonal groups of the short-finned squid,
Illex coindetii from the Strait of Sicily (central Mediterranean).
ICES Journal of Marine Science 57:3141.
Arkhipkin AI, Laptikhovsky VV. 2000. Age and growth of the
squid Todaropsis eblanae (Cephalopoda: Ommastrephidae) on
the north-west African shelf. Journal of the Marine Biological
Association of the UK 80:747 8.
Arkhipkin AI, Laptikhovsky VV, Golub A. 1999. Population
structure and growth of the squid Todarodes sagittatus (Cepha-
lopoda: Ommastrephidade) in the north-west African waters.
Journal of the Marine Biological Association of the UK
79:46777.
Bertalanffy L von. 1938. A quantitative theory of organic growth.
Human Biology 10:181 213.
Boyle PR, Ngoile MAK. 1993. Assessment of maturity state and
seasonality of reproduction in Loligo forbesi (Cephalopoda:
Loligo vulgaris growth strategies 57

Downloaded By: [University of Aberdeen] At: 15:42 2 March 2007
Loliginidae) from Scottish waters. In: Okutani T, O’Dor RK,
Kubodera T, editors. Recent Advances in Fishery Biology.
Tokyo: Tokai University Press. p 37 48.
Challier L, Orr P, Robin J-P. 2006a. Introducing inter-individual
growth variability in the assessment of a cephalopod popula-
tion: application to the English Channel squid Loligo forbesi .
Oecologia 150:1728.
Challier L, Pierce GJ, Robin J-P. 2006b. Spatial and temporal
variation in age and growth in juvenile Loligo forbesi and
relationships with recruitment in the English Channel and
Scottish waters. Journal of Sea Research 55:217 29.
Chen C, Chiu T. 2003. Variations on life history parameters in
two geographical groups of neon flying squid, Ommastrephes
bartramii , from the North Pacific. Fisheries Research 63:349
66.
Chen CS, Pierce GJ, Wang J, Robin J-P, Poulard JC, Pereira J, et
al. 2006. The apparent disappearance of Loligo forbesi from the
south of its range in the 1990s: trends in Loligo spp. abundance
in the northeast Atlantic and possible environmental influences.
Fisheries Research 78:44 54.
Cunha ME. 1993. Variabilidade estacional do zooplaˆncton na
plataforma continental Portuguesa [Seasonal variability of
zooplankton in the Portuguese mainland coast]. Boletim
UCA 1:22941.
Cunha MM, Moreno A, Pereira JMF. 1995. Spatial and temporal
occurrences of Loligo spp. in Portuguese waters. ICES. CM
1995/K:33.
Forsythe JW. 1993. A working hypothesis of how seasonal
temperature change may impact the field growth of young
cephalopods. In: Okutani T, O’Dor RK, Kubodera T, editors.
Recent Advances in Fishery Biology. Tokyo: Tokai University
Press. p 133 43.
Forsythe JW. 2004. Accounting for the effect of temperature on
squid growth in nature: from hypothesis to practice. Marine
Freshwater Research 55:331 9.
Forsythe JW, Van Heukelem WF. 1987. Growth. In: Boyle PR,
editor. Cephalopod Life Cycles, Vol. II: Comparative Reviews.
London: Academic Press. p 13556.
Grist EPM, des Clers S. 1999. Seasonal and genotypic influences
on life cycle synchronisation: further insights from annual
squid. Ecological Modelling 115:149 63.
Hatfield E. 2000. Do some like it hot? Temperature as a possible
determinant of variability in the growth of the Patagonian
squid, Loligo gahi (Cephalopoda: Loliginidae). Fisheries Re-
search 47:2740.
Hendrickson LC. 2004. Population biology of northern shortfin
squid (Illex illecebrosus ) in the nor thwest Atlantic Ocean and
initial documentation of a spawning area. ICES Journal of
Marine Science 61:25266.
Jackson GD. 1994. Application and future potential of statolith
increment analysis in squids and sepiods. Canadian Journal of
Fisheries and Aquatic Sciences 51:2612 25.
Jackson GD, Domeier ML. 2003. The effects of an extraordinary
El Nin
˜o/La Nin
˜a event on the size and growth of the squid
Loligo opalescens off Southern California. Marine Biology
142:92535.
Jackson GD, Forsythe JW, Hixon RF, Hanlon RT. 1997. Age,
growth, and maturation of Lolliguncula brevis (Cephalapoda:
Loliginidae) in the northwestern Gulf of Mexico with a
comparison of length-frequency versus statolith age analysis.
Canadian Journal of Fisheries and Aquatic Sciences 54:2907 
19.
Jackson GD, Moltschaniwskyj NA. 2002. Spatial and temporal
variation in growth rates and maturity in the Indo-Pacific squid
Sepioteuthis lessoniana (Cephalopoda: Loliginidade). Marine
Biology 140:74754.
Macy KW, Brodziak JKT. 2001. Seasonal maturity and size at age
of Loligo pealeii in waters of southern New England. ICES
Journal of Marine Science 58:852 64.
Markaida U, Quin
˜o´ nez-Velasquez C, Sosa-Nishizaki O. 2004.
Age, growth and maturation of jumbo squid Dosidicus gigas
(Cephalopoda: Ommastrephidae) from the Gulf of California,
Mexico. Fisheries Research 66:31 47.
Moltschaniwskyj NA. 1994. Muscle tissue and muscle fiber
dynamics in the tropical Loliginid squid Photololigo sp. (Ce-
phalopoda: Loliginidade). Canadian Journal of Fisheries and
Aquatic Sciences 51:830 5.
Moreno A. 1998. Variac
¸a
˜o sazonal na distribuic
¸a
˜o e abundaˆ ncia
dos cefalo´ podes da plataforma continental entre Espinho e
Nazare´ : resultados dos cruzeiros de Agosto e Novembro de
1996 no N/I ‘‘Mestre Costeiro’’. Relato´ rios Cientı´ficos e
Te´cnicos do Instituto de Investigac
¸a
˜o Pescas e do Mar 29.
Moreno A. 2002. Morfologia e micro-estrutura dos estato´ litos de
lula, Loligo vulgaris : Metodologias de determinac
¸a
˜o de idades.
Relato´ rios Cientı´ficos e Te´ cnicos do Instituto de Investigac
¸a
˜o
das Pescas e do Mar 86:1 46.
Moreno A, Pereira J, Arvanitidis C, Robin J-P, Koutsoubas D,
Perales-Raya C, et al. 2002. Biological variation of Loligo
vulgaris (Cephalopoda: Loliginidae) in the eastern Atlantic
and Mediterranean. Bulletin of Marine Science 71:515 34.
Moreno A, Pereira J, Cunha MM. 2005. Environmental influ-
ences on age and size-at-maturity of Loligo vulgaris. Aquatic
Living Resources 18:377 84.
Natsukari Y, Komine N. 1992. Age and growth estimation of the
European squid Loligo vulgaris , based on statolith microstruc-
ture. Journal of the Marine Biological Association of the UK
72:27180.
Natsukari Y, Nakanose T, Oda K. 1988. Age and growth of the
loliginid squid Photololigo edulis (Hoyle 1885). Journal of
Experimental Marine Biology and Ecology 116:17790.
Pecl GT. 2004. The in situ relationships between season of
hatching, growth and condition in the southern calamary,
Sepioteuthis australis. Marine and Freshwater Research 55:429
38.
Pecl GT, Moltschaniwskyj NA, Tracey S, Jordan A. 2004. Inter-
annual plasticity of squid life-history and population structure:
ecological and management implications. Oecologia 139:515 
24.
Pierce GJ, Boyle PR. 2003. Empirical modelling of interannual
trends in abundance of squid (Loligo forbesi) in Scottish waters.
Fisheries Research 59:305 26.
Pierce GJ, Boyle PR, Hastie LC, Santos MB. 1994a. Diets of
squid Loligo forbesi and Loligo vulgaris in the Northeast
Atlantic. Fisheries Research 21:149 63.
Pierce GJ, Guerra A. 1994. Stock assessment methods used for
cephalopod fisheries. Fisheries Research 21:255 85.
Pierce GJ, Hastie LC, Guerra A, Thorpe RS, Howard FG,
Boyle PR. 1994b. Morphometric variation in Loligo forbesi and
Loligo vulgaris : regional, seasonal, sex, maturity and worker
differences. Fisheries Research 21:117 48.
R Development Core Team. 2006. R: a language and environment
for statistical computing R Foundation for Statistical Comput-
ing. (http:\\www.r-project.org).
Raya CP, Balguerias E, Fernandez-Nunez MM, Pierce GJ. 1999.
On reproduction and age of the squid Loligo vulgaris from the
Saharan Bank (north-west African coast). Journal of the
Marine Biological Association of the UK 79:111 20.
Reynolds RW, Smith TM. 1994. Improved global sea surface
temperature analyses using optimum interpolation. Journal of
Climate 7:92948.
Ricker WE. 1979. Growth rates and models. In: Hoar WS,
Randall DJ, Brett JR, editors. Fish Physiology, Vol. III:
58 A. Moreno et al.

Downloaded By: [University of Aberdeen] At: 15:42 2 March 2007
Biogenics and Growth. Orlando: Academic Press. p 677 
743.
Robin J-P, Denis V. 1999. Squid stock fluctuations and water
temperature: temporal analysis of English Channel Loliginidae.
Journal of Applied Ecology 36:101 10.
Rocha F, Guerra A. 1999. Age and growth of two sympatric squid
Loligo vulgaris and Loligo forbesi , in Galician waters (north-west
Spain). Journal of the Marine Biological Association of the UK
79:697707.
Royer J, Peries P, Robin J-P. 2002. Stock assessments of English
Channel Loliginid squid: updated depletion methods and new
analytical methods. ICES Journal of Marine Science 59:445 
57.
Smith JM, Pierce GJ, Zuur AF, Boyle PR. 2005. Seasonal patterns
of investment in reproductive and somatic tissues in the squid
Loligo forbesi. Aquatic Living Resources 18:341 51.
Stearns SC. 1992. The Evolution of Life Histories. New York:
Oxford University Press.
Steer BLM, Jackson GD. 2004. Temporal shifts in the allocation
of energy in the arrow squid, Nototodarus gouldi : sex-specific
responses. Marine Biology 144:1141 9.
Turk PE, Hanlon RT, Bradford LA, Yang WT. 1986. Aspects of
feeding, growth and survival of the European squid Loligo
vulgaris Lamarck 1799, reared through the early growth stages.
Vie et Milieu 36:9 13.
Villanueva R. 2000. Effect of temperature on statolith growth of
the European squid Loligo vulgaris during early life. Marine
Biology 136:44960.
Villegas P. 2001. Growth, life cycle and fisheries biology of Loligo
gahi (d’Orbigny, 1835) off the Peruvian coast. Fisheries
Research 54:12331.
Young IAG, Pierce GJ, Dali HI, Santos MB, Key LN, Bailey N, et
al. 2004. Application of depletion methods to estimate stock
size in the squid Loligo forbesi in Scottish waters (UK). Fisheries
Research 69:21127.
Editorial responsibility: Alan Jaylor
Loligo vulgaris growth strategies 59