Abstract

González, Á. F., Otero, J., Pierce, G. J., and Guerra, Á. 2010. Age, growth, and mortality of Loligo vulgaris wild paralarvae: implications for understanding of the life cycle and longevity. – ICES Journal of Marine Science, 67: 1119–1127.

Age, growth, and mortality were estimated for the first time in wild paralarvae of the common squid, Loligo vulgaris, by examining growth increments in the statoliths of 273 animals collected off the Ría de Vigo (NW Spain, NE Atlantic). Hatching was all year round for the period 2003–2005, with a peak during late spring and a secondary peak during early autumn. Paralarvae varied from 1260 to 7580 µm, and their abundance decreased abruptly as they grew. Statolith increments were clearly visible without grinding in almost all material, allowing reliable estimation of age. Paralarvae are planktonic for at least 3 months. Growth in dorsal mantle length (DML) during that period fitted an exponential equation. The instantaneous relative growth rates were 2.11, 2.15, and 1.82% DML d−1 for 2003, 2004, and 2005, respectively, and there were no significant differences in size-at-age between the 3 years. Taking into account the growth rates estimated for the whole cycle of L. vulgaris, we suggest that the lifespan may previously have been underestimated by 3 months, because the proximity of the rings deposited during paralarval and early juvenile stages would prevent accuracy in enumerating the number of growth increments in later stages. The estimated instantaneous rate of total mortality during the first 90 d of a paralarva life was 9.6, 5.3, and 4.8% d−1 for 2003, 2004, and 2005, respectively. Eye diameter was a reliable and rapid way of estimating DML and age.

Introduction

Although the planktonic paralarvae of many cephalopods are relatively large, physically resemble the adults, and are capable of relatively rapid movement (Boletzky, 1974), their presence in plankton samples is rare (Boyle and Rodhouse, 2005). This could be because of the inadequacy of the sampling methods, and possibly also their patchy distribution (Piatkowski, 1998; Collins et al., 2002; González et al., 2005). Another uncertainty relates to the difficulty of identifying paralarvae to species (Hanlon et al., 1992). Poor sampling of cephalopod paralarvae has at least two negative consequences. First, the lack of information may preclude correct interpretation of the adult life cycle and its place in the structure and function of marine ecosystems (Vecchione, 1987). Second, its paucity explains why cephalopod paralarva surveys are not used widely for fisheries assessment purposes (see Boyle and Rodhouse, 2005, for a review).

Knowledge of recruitment to a fishery is particularly important in short-lived species, such as most cephalopods, in which there is a complete turnover of biomass every 1 or 2 years (Guerra, 2006). Recruitment success is related to both biotic factors and environmental conditions. Two recent studies undertaken on Octopus vulgaris emphasized the importance of studying the influence of oceanographic parameters on the spawning strategy and planktonic larval ecology to understand the natural variability in recruitment events (Otero et al., 2008, 2009).

The common squid (Loligo vulgaris) is a fast-growing cephalopod that inhabits temperate waters of the eastern Atlantic from the North Sea and British Isles to northern Namibia, including the Mediterranean Sea (Guerra, 1992). Although many studies have been carried out on subadult and adult squid throughout their range (see Boyle, 1983, and Boyle and Rodhouse, 2005, for reviews), and particularly in Galician waters (Guerra and Rocha, 1994; Rocha and Guerra, 1999), little is known of the abundance, distribution, age, growth, and mortality of the species' early stages, except for a few data on abundance and distribution of wild paralarvae off Galicia (Rocha et al., 1999; González et al., 2005).

In laboratory experiments, daily increment deposition in L. vulgaris statoliths was validated in paralarvae (Villanueva, 2000a). The effect of temperature on embryonic and post-hatching growth and on statolith increment-deposition rate was shown to be important in the species (Villanueva, 2000a, b; Villanueva et al., 2003). Besides those studies on early stages of development, daily increment deposition was also validated in cultured juveniles and adults of the subspecies L. vulgaris reynaudii (Durholtz et al., 2002), but a study on wild animals of the same subspecies showed that daily deposition can only be applied to males of 290–370 mm dorsal mantle length (DML; Lipiński et al., 1998).

To date, estimates of the age and growth of wild L. vulgaris populations have been based on reading the daily growth increments in the statoliths of subadults and adults (Natsukari and Komine, 1992; Arkhipkin, 1995; Bettencourt et al., 1996; Raya et al., 1999; Rocha and Guerra, 1999). The absence of information on paralarvae in these studies could represent an important bias in the interpretation of the real age and growth rate of L. vulgaris, because the increments deposited during the early stages of development are more difficult to read in larger statoliths.

Accurate estimate of mortality rates is difficult in short-lived species, and a large proportion of the estimates currently used in stock assessment is based on empirical relationships originally developed for fish, and they are applied to an unspecified part of the life history of some species (Caddy, 1996). Survival under controlled conditions has been reported for several loliginids (Yang et al., 1986; Hanlon et al., 1989; Villanueva, 2000a). However, Bigelow (1992), who estimated mortality based on growth increments in the statoliths of a few wild paralarvae of Abralia trigonura, noted the complete lack of estimates of mortality from field data, and this situation remains today.

The aim of this work was to assess the use of statolith microstructures in studying some demographic parameters of wild L. vulgaris paralarvae and to evaluate the implication of the results in relation to understanding the demographics of populations of post-paralarvae. Paralarval growth in length and weight per year is described, and changes in some morphometric characters with growth, hatching season, and mortality in different year classes of paralarvae are demonstrated.

Material and methods

Collection of wild paralarvae

In all, 47 biological surveys were undertaken onboard the RV “Mytilus” in Galician waters, NW Spain (Figure 1), from January 2003 to October 2005. The surveys varied in periodicity between years, but methodology was otherwise consistent. Plankton sampling was undertaken monthly during 2003, fortnightly between May and October 2004, and twice per week in July and late September and early October 2005. Four transects of 2.8 km were covered in each survey, average bottom depths ranging from 26 to 85 m (Figure 1). Owing to the small number of cephalopod paralarvae obtained from the inner transect (T1; Figure 1), during each of the monthly surveys of 2003, that transect was substituted by a deeper one (110 m, T5) in 2004 and 2005. Zooplankton samples were collected by towing, near-bottom and at the surface, a bongo net of 750 mm diameter equipped with 375 µm mesh. At a ship's speed of 2 knots, the bongo net was first lowered and stabilized near the bottom for a period of 15 min, then hauled to the surface at 0.5 m s−1. It was then redeployed to collect samples in surface waters only. The bongo was equipped with a current meter to allow calculation of the volume of water filtered during each haul, so permitting paralarval abundance to be estimated (as number per 1000 m3), along with a depth meter to help identify the water strata sampled during each haul.

Figure 1.

Map of the study area showing the four plankton transects performed in 2003 with average bottom depths ranging from 26 to 85 m. Because there were few cephalopod paralarvae taken on the inner transect (T1) in 2003, that transect was substituted by a deeper one (110 m, T5) in 2004 and 2005.

Figure 1.

Map of the study area showing the four plankton transects performed in 2003 with average bottom depths ranging from 26 to 85 m. Because there were few cephalopod paralarvae taken on the inner transect (T1) in 2003, that transect was substituted by a deeper one (110 m, T5) in 2004 and 2005.

Zooplankton samples were fixed on board with 4% buffered formalin, then transferred after 24 h to 70% ethanol. Paralarvae of L. vulgaris were separated and later identified in the laboratory according to the descriptions of Fioroni (1965) and Hanlon et al. (1992), and reference collections of L. vulgaris paralarvae hatched under rearing conditions.

Measurements

In all, 376 paralarvae were measured, but those damaged during collection (n = 9) were discarded from further study. Wet body weight (BW) was obtained to the nearest 0.1 mg using a Sartorius MC 210P precision balance, and with a Nikon SMZ 800 stereomicroscope, the following measurements were made to the nearest 0.01 µm, following Roper and Voss (1983): total length (TL), DML, ventral mantle length (VML), mantle width (MW), eye diameter (ED), longest arm length (AL), and tentacle length. Linear, power, exponential, and logarithmic equations were fitted to the relationships between DML and the other morphometric parameters. The same equations were also fitted to DML and BW, but only the best fit is shown here.

Age determination

The mantle and funnel of the paralarvae were removed to access the statoliths, which were visible as opaque structures within the statocysts. Statoliths were removed with fine dissecting needles (20 µm tip diameter) under a stereomicroscope. All statoliths were measured (to 0.01 µm) following the terminology of Clarke (1978), from the end of the dorsal dome to the tip of the rostrum (statolith length), and across the widest part of the dorsal dome (statolith width). The method applied to determine the age of paralarvae involved mounting the statolith on a microscope slide with Crystalbond, with the anterior concave side uppermost. The growth increments of most statoliths were clear because of their relative transparency. In a few cases, particularly the oldest paralarvae, the statoliths had to be ground, first on the anterior surface, then on the posterior surface. This grinding of both surfaces in the sagittal plane resulted in the production of relatively thin statolith sections. Increments were determined along the axis of maximum statolith growth with a NIS Elements D 2.30 image analysis system interfaced with a Nikon compound microscope (×400 magnification). Counts were obtained semi-automatically: putative increments were detected automatically by computer software from an enhanced image, but final identification of increments was carried out manually. In a few of the larger paralarvae, increments were not clear near the outer margin of the ground surface, so in those cases, the number of increments missed was estimated by extrapolation from the adjacent area (González et al., 2000). An age–length (DML) key was estimated for each year. Hatch date was back-calculated from the date of capture and the age of each paralarva.

Growth and mortality

Instantaneous relative growth rate (G, % DML d−1) for each year was calculated, using only the animals for which the age was estimated from daily growth increments on the statolith, following Forsythe and Van Heukelem (1987) as 

formula
where DML is the dorsal mantle length (µm) at time t (d). To estimate G for a given year, we calculated the average DML of paralarvae younger than 10 d (DML1) and the average DML of the oldest paralarvae, whose ages were within the final 10-d interval, e.g. 70–80 d old (DML2).

The instantaneous rate of total mortality (Z) for the years 2003, 2004, and 2005 was calculated using simple catch curves (Ricker, 1975). The paralarvae collected were grouped into age classes of equal breadth (10 d), and natural logarithms of the frequency of occurrence were plotted against the age classes. We chose the day as the unit of time to express mortality rates, following Caddy (1996), who indicated that although it is traditional in stock assessment, it is obvious that it is not practical to express instantaneous rates of mortality of short-lived species yearly.

Statistical analysis

Differences between years in DML vs. weight, DML vs. age, and survivorship vs. age were analysed using generalized additive models (Hastie and Tibshirani, 1990). For DML vs. weight, both variables were log-transformed, because the underlying relationship was expected to approximate a power function. In the other two cases, however, a better approximation to a Gaussian distribution and homogeneity of variance was achieved by log-transforming the response variable. Differences between years were determined by fitting separate smoothers (for the effect of the main explanatory variable, i.e. DML or age) for each year, as well as including year as a factor. That model was then compared with a model with a common smoother for all 3 years, using an F-test. The approach was equivalent to but more robust than the option of including linear interaction terms. As sampling months differed between years, models of weight-at-length and length-at-age that included month or season (April–September vs. October–March) as an explanatory variable were also explored (this was not possible for the analysis of survivorship because, for each year, it was necessary to use combined data from all months). For models of weight-at-length and length-at-age that included season, it was also possible to explore interactions between the effect of season and the effect of length and age, respectively. Moreover, because the conditions experienced around the time of hatching may be critical, we also repeated the analysis using hatch month instead of catch month as an explanatory variable. For the survivorship model, data were the numbers of animals surviving to a given age, and sample size was insufficient to make separate calculations for each month. All GAMs were fitted using BRODGAR software (Zuur et al., 2007).

Results

Correlates of growth of wild paralarvae

In all, 385 L. vulgaris paralarvae were collected during the 3-year sampling period. Their DML varied from 1260 to 7580 µm for the whole period studied. Of the total, 73% were small paralarvae ranging from 1500 to 3000 µm DML. Abundance decreased with increasing size once the animals reached around 2000 µm DML (Figure 2), and ∼5% of the largest paralarvae (>4000 µm DML) were collected at the deepest station (T5).

Figure 2.

DML (µm) distribution of the L. vulgaris paralarvae collected from 2003 to 2005.

Figure 2.

DML (µm) distribution of the L. vulgaris paralarvae collected from 2003 to 2005.

The relationships between DML and the five morphometric characters of paralarvae measured were all linear, and revealed high determination coefficients (Table 1), the highest value being for ED.

Table 1.

Equations of the relationships between L. vulgaris DML (µm) and the other measurements (µm) of the paralarvae (n = 376).

Equation Fit r2 
ED = 0.198 DML − 47.19 Linear 0.907 
VML = 0.812 DML − 174.14 Linear 0.902 
TL = 0.531 DML − 256.19 Linear 0.872 
AL = 0.328 DML − 255.14 Linear 0.850 
MW = 0.445 DML − 334.15 Linear 0.778 
Equation Fit r2 
ED = 0.198 DML − 47.19 Linear 0.907 
VML = 0.812 DML − 174.14 Linear 0.902 
TL = 0.531 DML − 256.19 Linear 0.872 
AL = 0.328 DML − 255.14 Linear 0.850 
MW = 0.445 DML − 334.15 Linear 0.778 

Figure 3 illustrates the DML–BW relationships for the 3 years of sampling. Initial exploration of GAMs including month as a continuous explanatory variable indicated that differences between months were not significant. However, length–weight relationships differed significantly between years, i.e. there was a significant interaction between the effects of year and DML, and the model with separate smoothers for the effect of DML on weight in each year was a significant improvement on the model with a common smoother. For all 3 years, the relationship between log-transformed weight and log-transformed DML was close to linear (Table 2). The fitted values from the GAM showed that the most obvious difference was between 2004 and 2005, animals being heavier in 2005 than in 2004.

Table 2.

GAM results for L. vulgaris BW in relation to DML and year.

Explanatory variable Fit Coefficient or d.f. Statistic Probability 
Year 2004 Linear 0.0323 t = 2.075 0.039 
Year 2005 Linear 0.0351 t = 2.095 0.037 
DML year 2003 Smoother d.f. 1.44 F = 213.1 <0.0001 
DML year 2004 Smoother d.f. 1.00 F = 1 508.9 <0.0001 
DML year 2005 Smoother d.f. 1.95 F = 552.8 <0.0001 
Explanatory variable Fit Coefficient or d.f. Statistic Probability 
Year 2004 Linear 0.0323 t = 2.075 0.039 
Year 2005 Linear 0.0351 t = 2.095 0.037 
DML year 2003 Smoother d.f. 1.44 F = 213.1 <0.0001 
DML year 2004 Smoother d.f. 1.00 F = 1 508.9 <0.0001 
DML year 2005 Smoother d.f. 1.95 F = 552.8 <0.0001 

BW and DML were log-transformed and a Gaussian GAM fitted (n = 376). The main effect of year is presented as comparisons of levels 2 and 3 (years 2004 and 2005) with level 1 (year 2003). Year also has a significant interaction with DML, such that the model with separate smoother terms for DML in each year was significantly better than the model with a single smoother for DML (F = 4.99, p = 0.01). The final model explains 92.5% of the deviance.

Figure 3.

DML (µm) of L. vulgaris plotted against BW (µg) for the 3 years of sampling; 2003, plus signs; 2004, squares; 2005, triangles.

Figure 3.

DML (µm) of L. vulgaris plotted against BW (µg) for the 3 years of sampling; 2003, plus signs; 2004, squares; 2005, triangles.

Age and growth

Statolith increments were clear without grinding in almost all paralarvae, allowing reliable estimation of age (Figure 4). Statoliths from 273 paralarvae, for which the DML ranged from 1400 to 7580 µm, were read. An exponential model was the best fit to the growth of the paralarvae up to 80 d old (Figure 5a). The best estimates of instantaneous relative growth rate (G) for the wild paralarvae of L. vulgaris of that size range in Galician waters were 2.11, 2.15, and 1.82% DML d−1 for 2003, 2004, and 2005, respectively. ED was determined to be a reliable parameter to estimate age (n = 273; r2 = 0.80).

Figure 4.

Light micrograph of a statolith from a 1.9-mm DML L. vulgaris paralarva. Growth increments (d) are clearly visible without grinding. The hatching increment is shown.

Figure 4.

Light micrograph of a statolith from a 1.9-mm DML L. vulgaris paralarva. Growth increments (d) are clearly visible without grinding. The hatching increment is shown.

Figure 5.

DML (µm) of L. vulgaris plotted against age (d) for the period 2003–2005; 2003 squares; 2004, circles; 2005, plus signs.

Figure 5.

DML (µm) of L. vulgaris plotted against age (d) for the period 2003–2005; 2003 squares; 2004, circles; 2005, plus signs.

Initial GAM fits revealed a marginally significant trend of length-at-age being smaller later in the year, so month was retained in the model. The final model included a weak negative effect of month, but no significant interannual variation (Table 3). Inclusion of hatch month rather than catch month in the model resulted in almost no change in the overall model; the significance of hatch month was p = 0.032, compared with p = 0.047 for catch month.

Table 3.

GAM results for L. vulgaris DML in relation to age and month.

Explanatory variable Fit Coefficient or d.f. Statistic Probability 
Month Linear −0.0025 t = −1.995 0.047 
Age Smoother d.f. 2.85 F = 645.6 <0.0001 
Explanatory variable Fit Coefficient or d.f. Statistic Probability 
Month Linear −0.0025 t = −1.995 0.047 
Age Smoother d.f. 2.85 F = 645.6 <0.0001 

DML was log-transformed to improve normality and a Gaussian GAM was fitted (n = 271). The model explained 88.9% of the deviance.

Classifying the months as spring/summer (April–September) and autumn/winter (October–March), there was a significant interaction between the effects of age and season (p = 0.004), i.e. the length–age relationship was less linear in spring and summer, although the main effect of season was not significant. If the year was similarly divided according to hatch month rather than catch month, there was no such interaction.

Hatching season

Figure 6a shows the abundance of L. vulgaris paralarvae collected from 2003 to 2005. Data on catch day and estimation of age allowed us to infer the date of hatching of our L. vulgaris paralarvae. Hatching was year-round, with a peak in late spring and early summer, and a secondary peak in early autumn. The older paralarvae taken in September would have hatched during the main peak, and the young ones collected in September or October and the old ones collected from October to December would have been derived from the secondary hatching peak (Figure 6b). This finding suggests relatively less hatching during winter.

Figure 6.

(a) Hatching season for L. vulgaris paralarvae based on mean abundance (number of individuals per 1000 m3) for the period 2003–2005. (b) Monthly mean age of paralarvae collected from 2003 to 2005.

Figure 6.

(a) Hatching season for L. vulgaris paralarvae based on mean abundance (number of individuals per 1000 m3) for the period 2003–2005. (b) Monthly mean age of paralarvae collected from 2003 to 2005.

Mortality of planktonic paralarvae

Figure 7 depicts the number of paralarvae per age class (10 d intervals) and the catch curves for the years 2003–2005. The instantaneous rate of total mortality for the first 90 d of life was 9.6, 5.3 and 4.8% d−1 for 2003, 2004, and 2005, respectively. GAM results indicated that survivorship at age was better in 2004 than in 2003 and that there was also a significant interaction between year and age effects. Comparison of the shapes of the smoothers for each year suggested that the main difference was in survival up to the age of 60 d (Table 4; Figure 8). Note, however, that it was not possible to include month as a factor in this analysis, so differences between years may have been affected by the seasonal composition of samples.

Table 4.

GAM results for L. vulgaris survivorship in relation to age and year.

Explanatory variable Fit Coefficient or d.f. Statistic Probability 
Year 2004 Linear 0.3021 t = 3.877 0.0009 
Year 2005 Linear −0.2011 t = 2.581 0.0175 
Age year 2003 Smoother d.f. 2.34 F = 50.00 <0.0001 
Age year 2004 Smoother d.f. 2.83 F = 38.38 <0.0001 
Age year 2005 Smoother d.f. 1.00 F = 85.95 <0.0001 
Explanatory variable Fit Coefficient or d.f. Statistic Probability 
Year 2004 Linear 0.3021 t = 3.877 0.0009 
Year 2005 Linear −0.2011 t = 2.581 0.0175 
Age year 2003 Smoother d.f. 2.34 F = 50.00 <0.0001 
Age year 2004 Smoother d.f. 2.83 F = 38.38 <0.0001 
Age year 2005 Smoother d.f. 1.00 F = 85.95 <0.0001 

Survivorship was log-transformed to improve normality and a Gaussian GAM was fitted (n = 30). The model with separate smoother terms for age in each year explained 94.6% of deviance and was significantly better than the model with a single smoother for age (F = 5.09, p = 0.006).

Figure 7.

Logarithms of number of L. vulgaris wild paralarvae of successive age periods (d) in samples from the Ria de Vigo. The catch curve equations for each sampling year are given. The instantaneous rate of total mortality corresponds to the slope of the regressions.

Figure 7.

Logarithms of number of L. vulgaris wild paralarvae of successive age periods (d) in samples from the Ria de Vigo. The catch curve equations for each sampling year are given. The instantaneous rate of total mortality corresponds to the slope of the regressions.

Figure 8.

Smoothers for the partial effect of age (d) on survivorship in each year. Survivorship was log-transformed. Dashed lines indicate 95% confidence limits: (a) 2003, (b) 2004, and (c) 2005. Labelling of the y-axis follows a standard format whereby s(X, edf) indicates the effect of the explanatory variable X when expressed as a smooth function with an estimated degrees of freedom (edf).

Figure 8.

Smoothers for the partial effect of age (d) on survivorship in each year. Survivorship was log-transformed. Dashed lines indicate 95% confidence limits: (a) 2003, (b) 2004, and (c) 2005. Labelling of the y-axis follows a standard format whereby s(X, edf) indicates the effect of the explanatory variable X when expressed as a smooth function with an estimated degrees of freedom (edf).

Discussion

The smallest hatching size of wild L. vulgaris paralarvae from Galicia was only around half the size of those hatching in the Mediterranean (Boletzky, 1979; Turk et al., 1986). We considered the possibility that the difference in size could be explained if the smaller paralarvae were Alloteuthis spp., the only other loliginid present in our geographic area (Guerra, 1992), but the possibility was rejected because the loliginid paralarvae we collected had two rows of red chromatophores in the tentacles rather than one, as in Alloteuthis (Fioroni, 1965). The biological plasticity reported for L. vulgaris and other short-lived loliginids (Boyle and Rodhouse, 2005) could explain why paralarvae hatch at smaller sizes in Galician waters. Furthermore, rearing conditions can alter the size at hatching (Villanueva, 2000a), but there is currently little prospect of identifying the environmental factors that might account for this phenomenon.

Among the morphometric parameters measured in the paralarvae of L. vulgaris in Galician waters, DML and ED were closely correlated. As the eye is almost always intact in the paralarvae captured by nets, its measurement provides a reliable, accurate, and rapid way of estimating paralarva DML and age, especially when the mantle of paralarvae is damaged during capture.

According to our results, the paralarvae of L. vulgaris hatch throughout the year in Galician waters, with a peak in late spring and early summer, and a secondary peak in early autumn. This agrees with the results of studies undertaken by Moreno et al. (2009) in Portuguese waters and by Guerra and Rocha (1994), who observed that the reproductive period of the species in Galician waters extended throughout the year, with a season of intensive spawning from December to April. The latter would produce the peaks of hatching listed above.

Comparisons between several hard structures revealed that an analysis of growth increments in statoliths remains the best way to estimate the age of squid (González et al., 2000). Nevertheless, validation is necessary to confirm that the deposition of growth increments is daily, a premise that was demonstrated in cultured paralarvae of L. vulgaris by Villanueva (2000a) and, inter alia, in cultured juveniles and adults of the subspecies L. vulgaris reynaudii (Durholtz et al., 2002). In statoliths of both juveniles and adults of most species, there is an area close to the nucleus where increments cannot be clearly discerned because they are very close together, and there is a thick wing with amorphous crystallization. This could lead to underestimating adult age and hence introduce a bias into the interpretation of maturity and mortality data (González et al., 2000; Hendrickson and Hart, 2006). These issues underscore the importance of applying age-determination techniques to statoliths of wild squid paralarvae, because increments read in a paralarva statolith would subsequently become obscure as the squid enters the juvenile stage. The advantage of reading paralarval statoliths, at least during the first few months, is that grinding of this hard structure is generally not necessary, although reading remains a difficult and time-consuming process.

The first age estimates for wild L. vulgaris paralarvae, presented here, indicated that this species inhabits the plankton for about 3 months (up to 8000 µm DML) in Galician waters. The paralarvae gradually disappear from the mesozooplankton fraction as they grow, mainly as a consequence of heavy mortality during that period, and also probably because the survivors become more nektonic.

As far as we can ascertain, this is the first time that instantaneous relative growth rates (G) have been estimated for wild paralarvae of loliginid squid. The overall G for squid up to 3 months old, estimated from animals collected in Galician waters, is within the range of values obtained by Villanueva (2000a) in culture-based studies of paralarvae from the same species in the Mediterranean. Our results are also consistent with the range of values of G obtained by Turk et al. (1986), also using reared squid from the Mediterranean. The lower value of the range obtained by those authors (1.07) coincides with the results of Boletzky (1979), also for Mediterranean animals but reared at lower temperatures and with less variety and lower density of food. However, our data differ from the age and growth rates estimated from statolith analysis by Natsukari and Komine (1992) of older wild Mediterranean animals of >60 mm DML. This discrepancy could be explained because the G and the age of small squid (<60 mm DML) estimated by those authors were calculated using an exponential model fitted only to larger squid.

Rocha and Guerra (1999) estimated ages ranging from 167 to 382 d for L. vulgaris of 92–383 mm DML, with an estimated G of 0.53–0.84% DML d−1. If the growth rates estimated in the present work for the first 90 d of life of paralarvae (1.82–2.15) remained constant throughout the life of the animal, squid of 92 mm DML would attain that size in 191–226 d, and squid of 383 mm DML would reach that size in 257–305 d. However, as shown in other loliginid and oegopsid squid (Natsukari and Komine, 1992; González et al., 1996; Boyle and Rodhouse, 2005), the value of G decreases with age.

Assuming constant growth during the whole life cycle, the highest of the Rocha and Guerra (1999) values of G (0.84), which is close to the value (0.81) estimated here for the final part of the wild paralarva stage (age 70–90 d), squid would reach 92 mm DML after 306 d, and 383 mm DML after 460 d. These calculations suggest that Rocha and Guerra (1999) underestimated the age of juveniles and adult L. vulgaris by about 3 months. However, as G decreases with age (Natsukari and Komine, 1992; González et al., 1996; Boyle and Rodhouse, 2005), these underestimates of age and lifespan in adult squid could be higher and, in part, perhaps reflect the stated proximity of the rings deposited in statoliths during the paralarva stage, implying that this issue has not previously been taken into account adequately.

The results of laboratory studies (see Forsythe and Van Heukelem, 1987, for a review; Hatfield et al., 2001) have consistently demonstrated that the growth in BW of loliginids and benthic octopods takes place in two phases of the life cycle. The first phase is exponential in form, with a constant rate of growth of 4–8%, depending on species. The second phase is logarithmic and lasts until near the end of the squid's life. Our data on BW and age, fitted to an exponential equation as BW (µg) = 0.653 e0.0584Age; r2 = 0.834) yielded a G of 6.29% BW d−1 for the whole period sampled. These data agree with the revision made by Forsythe and Van Heukelem (1987). However, we prefer to use growth in DML, because the accuracy of that measure is much greater than that for BW. On the other hand, one of the most important variables affecting growth rates is temperature (Forsythe, 1993; Hatfield, 2000; Hatfield et al., 2001). This could be one of the explanations for the differences in growth rate (expressed in DML d−1) between the three sampling years here.

The estimates here of mortality for the planktonic period of L. vulgaris paralarvae are the first available for myopsid squid, and they agree with the estimate of Bigelow (1992) for the oegopsid squid A. trigonura, also based on growth increments in statoliths. In 2004 and 2005, larger paralarvae were collected than in 2003 (Figure 2), coinciding with sampling taking place at the deepest offshore station (T5) in 2004 and 2005 but not in 2003. This could be interpreted as larger paralarvae emigrating to deeper water, but this does not seem to have been the case, because <5% of the largest paralarvae were collected at the deepest offshore station. Therefore, the variability in paralarva catch curves could be interpreted also as variable mortality.

We compared length–weight and size-at-age relationships between years and between months. Although the statistical approach we used (GAM) should be reasonably robust to unbalanced sampling, ideally, data are needed for a wider range of months in all years. For survivorship-at-age, it is not possible to separate annual and seasonal effects. The analysis revealed interannual differences in length–weight relationships and survivorship, which might be expected to relate to external factors such as food availability. Conversely, length-at-age varied seasonally but not between years, suggesting that it relates primarily to seasonal patterns in growth rate, for instance in response to temperature differences.

Loligo vulgaris wild paralarvae remain in the plankton for ∼3 months, growing fast and with insignificant interannual variation. However, mortality differs significantly between years, potentially influenced by biotic and abiotic drivers. Studies on the ecology of wild paralarvae, and especially their relationship with physical and chemical oceanographic parameters, are scarce, so we encourage more, to advance the knowledge of the early stages of development, the critical point in a cephalopod's life cycle.

Acknowledgements

The research was funded by the CICYT (REN2002–02111/MAR and VEM2003–20010) and Xunta de Galicia (PGIDIT02–RMA–C40203PR) projects. JO was supported by fellowships of the “Diputación Provincial de Pontevedra” and the European Social Fund–CSIC I3P postgraduate programme. GJP acknowledges support from the EU through the ANIMATE project (MEXC-CT-2006-042337).We thank the personnel of the RV “Mytilus” for assistance in the field, and Mike Vecchione, an anonymous referee, and the editor for their valued comments on the submitted manuscript.

References

Arkhipkin
A.
Age, growth and maturation of the European squid Loligo vulgaris (Myopsida, Loliginidae) on the west Saharan shelf
Journal of the Marine Biological Association of the UK
 , 
1995
, vol. 
75
 (pg. 
592
-
604
)
Bettencourt
V.
Coelho
M. L.
Andrade
J. P.
Guerra
A.
Age and growth of the squid Loligo vulgaris off the south coast of Portugal, using statolith analysis
Journal of Molluscan Studies
 , 
1996
, vol. 
62
 (pg. 
359
-
366
)
Bigelow
K. A.
Age and growth of mesopelagic squid Abralia trigonura based on daily growth increments in statoliths
Marine Ecology Progress Series
 , 
1992
, vol. 
82
 (pg. 
31
-
40
)
Boletzky
S. v.
The ‘larvae’ of Cephalopoda. A review
Thalassia Jugoslavia
 , 
1974
, vol. 
10
 (pg. 
45
-
76
)
Boletzky
S. v.
Observations on the early post-embryonic development of Loligo vulgaris (Mollusca, Cephalopoda)
Rapport Commission Internationale Mer Méditerranée
 , 
1979
, vol. 
25/26
 (pg. 
155
-
158
)
Boyle
P. R.
Cephalopod Life Cycles. 1. Species Accounts
1983
London
Academic Press
Boyle
P. R.
Rodhouse
P.
Cephalopods. Ecology and Fisheries
2005
Oxford
Blackwell Science
pg. 
452
 
Caddy
J.
Modelling natural mortality with age in short-lived invertebrate populations: definition of a strategy of genomic time division
Aquatic Living Resources
 , 
1996
, vol. 
9
 (pg. 
197
-
207
)
Clarke
M. R.
The cephalopod statolith—an introduction to its form
Journal of the Marine Biological Association of the UK
 , 
1978
, vol. 
58
 (pg. 
701
-
712
)
Collins
M. A.
Yau
C.
Boyle
P. R.
Friese
D.
Piatkowski
U.
Distribution of cephalopods from plankton surveys around the British Isles
Bulletin of Marine Science
 , 
2002
, vol. 
71
 (pg. 
239
-
254
)
Durholtz
M. D.
Lipiński
M. R.
Field
J. G.
Laboratory validation of periodicity of incrementation in statoliths of the South African chokka squid Loligo vulgaris reynaudii (d'Orbigny, 1845): a re-evaluation
Journal of Experimental Marine Biology and Ecology
 , 
2002
, vol. 
279
 (pg. 
41
-
59
)
Fioroni
P.
Die embrionale musterentwicklung bei einigen Mediterranen tintenfishatren
Vie Milieu
 , 
1965
, vol. 
16
 (pg. 
656
-
756
)
Forsythe
J. W.
Okutani
T.
O'Dor
R. K.
Kubodera
T.
A working hypothesis of how seasonal temperature change may impact the field growth of young cephalopods
Recent Advances in Cephalopod Fisheries Biology
 , 
1993
Tokyo
Tokai University Press
(pg. 
133
-
144
)
Forsythe
J. W.
Van Heukelem
W. F.
Boyle
P. R.
Growth
Cephalopod Life Cycles. 2. Comparative Reviews
 , 
1987
London
Academic Press
(pg. 
135
-
156
)
González
Á. F.
Castro
B. G.
Guerra
Á.
Age and growth of the short-finned squid Illex coindetii in Galician waters (NW Spain) based on statolith analysis
ICES Journal of Marine Science
 , 
1996
, vol. 
53
 (pg. 
802
-
810
)
González
Á. F.
Dawe
E. G.
Beck
P.
Perez
J. A. A.
Bias associated with statolith-based methodologies for ageing squid; a comparative study on Illex illecebrosus (Cephalopoda: Ommastrephidae)
Journal of Experimental Marine Biology and Ecology
 , 
2000
, vol. 
244
 (pg. 
161
-
180
)
González
Á. F.
Otero
J.
Guerra
Á.
Prego
R.
Rocha
F.
Dale
A. W.
Distribution of common octopus and common squid paralarvae in a wind-driven upwelling area (Ria of Vigo, northwestern Spain)
Journal of Plankton Research
 , 
2005
, vol. 
27
 (pg. 
271
-
277
)
Guerra
Á.
Ramos
M. A.
Alba
J.
Bellés
X.
Gosálvez
J.
Guerra
A.
Macpherson
E.
Piera
F.
, et al.  . 
Mollusca, Cephalopoda
Fauna Ibérica, 1
 , 
1992
Madrid
Museo Nacional de Ciencias Naturales, CSIC
(pg. 
108
-
110
327
Guerra
Á.
Estrategia evolutiva de los cefalópodos
Investigación y Ciencia
 , 
2006
, vol. 
355
 (pg. 
50
-
59
)
Guerra
Á.
Rocha
F.
The life history of Loligo vulgaris and Loligo forbesi (Cephalopoda: Loliginidae) in Galician waters (NW Spain)
Fisheries Research
 , 
1994
, vol. 
21
 (pg. 
43
-
69
)
Hanlon
R. T.
Boletzky
S. v.
Okutani
T.
Pérez-Gándaras
G.
Sánchez
P.
Sousa-Reis
C.
Vecchione
M.
Suborder Myopsida Orbigny, 1845 (Order Teuthoidea Naef, 1916)
Smithsonian Contribution to Zoology
 , 
1992
, vol. 
513
 (pg. 
37
-
53
)
Hanlon
T.
Yang
W. T.
Turk
P. E.
Lee
P. G.
Hixon
R. F.
Laboratory culture and estimated life span of the eastern Atlantic squid, Loligo forbesi Steenstrup, 1856 (Mollusca: Cephalopoda)
Aquaculture and Fisheries Management
 , 
1989
, vol. 
20
 (pg. 
15
-
34
)
Hastie
T.
Tibshirani
R. J.
Generalized Additive Models
1990
London
Chapman and Hall/CRC Press
Hatfield
E. M. C.
Do some like it hot? Temperature as a possible determinant of variability in the growth of the Patagonian squid, Loligo gahi (Cephalopoda: Loliginidae)
Fisheries Research
 , 
2000
, vol. 
47
 (pg. 
27
-
40
)
Hatfield
E. M. C.
Hanlon
R. T.
Forsythe
J. W.
Grist
E. P. M.
Laboratory testing of a growth hypothesis for juvenile squid, Loligo pealeii (Cephalopoda: Loliginidae)
Canadian Journal of Fisheries and Aquatic Sciences
 , 
2001
, vol. 
58
 (pg. 
845
-
857
)
Hendrickson
L. D.
Hart
D. R.
An age-based cohort model for estimating the spawning mortality of semelparous cephalopods with an application to per-recruit calculations for the northern shortfin squid, Illex illecebrosus
Fisheries Research
 , 
2006
, vol. 
78
 (pg. 
4
-
13
)
Lipiński
M. R.
Durholtz
M. D.
Underhill
L. G.
Field validation of age readings from the statoliths of chokka squid (Loligo vulgaris reynaudii d'Orbigny, 1845) and an assessment of associated errors
ICES Journal of Marine Science
 , 
1998
, vol. 
55
 (pg. 
258
-
270
)
Moreno
A.
Dos Santos
A.
Piatkowski
U.
Santos
A. M. P.
Cabral
H.
Distribution of cephalopod paralarvae in relation to the regional oceanography of western Iberia
Journal of Plankton Research
 , 
2009
, vol. 
31
 (pg. 
73
-
91
)
Natsukari
Y.
Komine
N.
Age and growth estimation of the European squid Loligo vulgaris, based on statolith microstructures
Journal of the Marine Biological Association of the UK
 , 
1992
, vol. 
72
 (pg. 
271
-
280
)
Otero
J.
Salgado
X. A.
González
Á. F.
Gilcoto
M.
Guerra
Á.
High-frequency coastal upwelling events influence Octopus vulgaris larval dynamics on the NW Iberian shelf
Marine Ecology Progress Series
 , 
2009
, vol. 
386
 (pg. 
123
-
132
)
Otero
J.
Salgado
X. A.
González
Á. F.
Groom
S. B.
Miranda
A.
Cabanas
J. M.
Guerra
Á.
Bottom-up and top-down controls of common octopus (Octopus vulgaris) in a wind-driven upwelling area (Galicia, NE Atlantic)
Marine Ecology Progress Series
 , 
2008
, vol. 
362
 (pg. 
181
-
192
)
Piatkowski
U.
Modern target sampling techniques provide new insights into the biology of early life stages of pelagic cephalopods
Biologie Mar Méditerranée
 , 
1998
, vol. 
5
 (pg. 
260
-
272
)
Raya
C. P.
Balguerías
E.
Fernández-Núñez
M. M.
Pierce
G. J.
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
 , 
1999
, vol. 
79
 (pg. 
111
-
120
)
Ricker
W. E.
Computation and interpretation of biological statistics of fish populations
1975
Bulletin of the Fishery Research Board of Canada, 191
pg. 
382
 
Rocha
F.
Guerra
Á.
Age and growth of two sympatric squids Loligo vulgaris and Loligo forbesi, in Galician waters (NW Spain)
Journal of the Marine Biological Association of the UK
 , 
1999
, vol. 
79
 (pg. 
697
-
707
)
Rocha
F.
Guerra
Á.
Prego
R.
Piakowski
U.
Cephalopod paralarvae and upwelling conditions off Galician waters (NW Spain)
Journal of Plankton Research
 , 
1999
, vol. 
21
 (pg. 
21
-
33
)
Roper
C. F. E.
Voss
G. L.
Guidelines for taxonomic descriptions of cephalopod species
Memoirs of the National Museum of Victoria
 , 
1983
, vol. 
44
 (pg. 
49
-
63
)
Turk
P. E.
Hanlon
R. T.
Bradford
L. A.
Young
W. T.
Aspects of feeding, growth and survival of the European squid Loligo vulgaris Lamarck, 1979, reared through the early growth stages
Vie Milieu
 , 
1986
, vol. 
36
 (pg. 
9
-
13
)
Vecchione
M.
Boyle
P. R.
Juvenile ecology
Cephalopod Life Cycles. 2. Comparative Reviews
 , 
1987
London
Academic Press
(pg. 
61
-
84
)
Villanueva
R.
Effect of temperature on statolith growth of the European squid Loligo vulgaris during early life
Marine Biology
 , 
2000
, vol. 
136
 (pg. 
449
-
460
)
Villanueva
R.
Differential increment-deposition rate in embryonic statolith of the loliginid squid Loligo vulgaris
Marine Biology
 , 
2000
, vol. 
137
 (pg. 
161
-
168
)
Villanueva
R.
Arkhipkin
A. I.
Jereb
P.
Lefkaditou
E.
Lipiński
M. R.
Raya
C. P.
Riba
J.
, et al.  . 
Embryonic life of the loliginid squid Loligo vulgaris: comparison between statoliths of Atlantic and Mediterranean populations
Marine Ecology Progress Series
 , 
2003
, vol. 
253
 (pg. 
197
-
208
)
Yang
W. T.
Hixon
R. F.
Turk
P. E.
Krejci
M. E.
Hulet
W. H.
Hanlon
R. T.
Growth, behaviour and sexual maturation of the market squid Loligo opalescens cultured through the life cycle
Fishery Bulletin US
 , 
1986
, vol. 
84
 (pg. 
771
-
798
)
Zuur
A. F.
Ieno
E. N.
Smith
G. M.
Analysing Ecological Data. Statistics for Biology and Health Series
2007
New York
Springer
pg. 
685