Introduction

The bigfin reef squid Sepioteuthis lessoniana is a neritic squid widely distributed in Indo-West Pacific waters and regarded as a valuable fishery resource [1]. A species complex may exist in Australian waters based on genetic analysis [2], and Japanese waters based on various reproductive features and genetic analysis [3–5]. Nevertheless, recent study suggests a low genetic diversity for populations in the Japanese waters and significant separation for populations in Japanese waters from those in the East and South China Sea [6].

Sepioteuthis lessoniana occurs throughout the waters around Taiwan, usually caught as a by-catch of neritic trawl fisheries. The squid is also targeted by recreational fishers in the waters off northern Taiwan jigging from October through March. However, the fisheries biology and population structure of this squid around Taiwan remains poorly understood. The commercial value of S. lessoniana (ca. US$15 per kg; currency US$1 = NTD$30) is usually higher than other neritic squids (ca. US$4 per kg) [7]. Therefore, this squid is a marine resource with commercial potential in Taiwan. It is essential and critical to understand the life-history traits and population structure of S. lessoniana around Taiwan to develop approaches to conservation and management.

Squid statoliths are analogous to fish otoliths in structure and function [8], making them conveniently applicable to the techniques and methods which have been developed for otoliths [9, 10]. Enlargement over time and regular deposition make statoliths appropriate tissues for studying the age and growth of squids, and other life-history events which might be recorded in the statoliths [10]. Statoliths have been frequently used in age and growth studies of squids around the world [11–13].

The widespread distribution of S. lessoniana in neritic waters throughout the Indo-Pacific and those successfully maintained in aquaria make the species a useful model animal for exploring the influence of environmental conditions on growth and maturity [14–19]. Daily formation of growth increments within statoliths of S. lessoniana has been validated by chemical marking for populations in Australian waters [20]. Squids have been observed to grow rapidly and have short life spans with maximum age estimated to be ca. 200 days for populations in tropical waters (from Philippines [21] and from Thailand and Australia [17]). The spawning seasons of S. lessoniana ranged between April and October in Okinawa waters [22] and between July and November in Zanzibar waters [23]. Similar to other squids, growth of S. lessoniana was highly sensitive to the ambient environmental conditions such as water temperature [15, 17]. Therefore, growth and hence size-at-maturity may vary among populations experiencing different environmental conditions [18]. Study of life-history traits of a species throughout its distribution ranges is required to not only improve understanding of population parameters, but also to provide essential information for management [24].

The waters off northern Taiwan, located at the southern edge of the East China Sea, are influenced by the Taiwan Strait waters as well as the seasonal intrusion of the Kuroshrio current [25] resulting in abundant primary productivity [26], and thus is one of the important fishing grounds of Taiwan [27]. Growth and maturation are critical life-history traits in population ecology and are essential information for approaches to conservation and management. Sepioteuthis lessoniana occur throughout the waters around Taiwan as a by-catch of multi-gear fisheries, while relevant information on population biology and ecology are scarce. The objective of this study was to examine growth and maturation of S. lessoniana in the waters off northern Taiwan, where this squid is targeted by recreational fishing, using statolith microstructure analysis. The results will provide basic information and improve our understanding of S. lessoniana populations in the waters off northern Taiwan.

Materials and methods

Samples of S. lessoniana were collected monthly from a day-trip vessel which jigged squid with wood-shrimp bait around Keelung Islet in the waters off northern Taiwan between April 2009 and March 2010 (Fig. 1). The squid samples were frozen immediately then thawed before examination in the laboratory. Specimens were sexed and measurements were obtained of the dorsal mantle length (ML), body weight (BW), gonad organs weight (GW, oviducts and oviducal glands, nidamental glands, and ovaries in females; Needham complex and testes in males). Measurements were recorded to the nearest 1 mm and 1 g for length and weight measurements, respectively.

Fig. 1
figure 1

Sampling area (shaded) of Sepioteuthis lessoniana in the waters off northern Taiwan

Full size image

Sexual maturity was determined with the naked eye following Boyle and Rodhouse [28] classification: stage I and II (immature); III (maturing); IV (mature); and V (spent). A pair of statoliths were extracted from each specimen, washed and stored in 96-well plates with liquid paraffin.

Each statolith (left side) was cleaned with xylene and ethanol and mounted on a glass slide with crystal bond. A transverse plane was obtained by grinding from the anterior side. The statolith was ground by hand with wet waterproof sandpaper (1,000, 2,400 and 4,000 grits, Buehler), and polished with 0.05 µm alumina powder and polish cloth. The prepared statoliths were examined under a compound microscope (Leica DM-2500) with a digital camera (Canon EOS-500D). Frames of statolith images were taken in series from the nucleus to margin of the lateral dome and saved as digital image files. Multiple frames of a statolith image were stitched together as a whole using Photoshop software, and the growth increments of the statolith were then traced and counted using Powerpoint software. The total number of increments (NI) within each statolith was marked and counted from the nucleus to the tip of the lateral dome. One observer marked and counted the samples most of the time. However, all the increments were marked on prepared statolith images and saved as image files which are traceable if any sample was questionable.

The formation of growth increments in statoliths of S. lessoniana has been validated to be daily [20]. Accordingly, the number of increments was considered to be equivalent to the age of the squid in days. Hatching dates were determined by subtracting estimated daily ages from the date of capture.

The relationship between ML and BW was fitted using a power function: BW = aMLb, where a and b are coefficients to be estimated. The effect of sex on the exponent b of the ML-BW relationship was examined by analysis of covariance (ANCOVA).

Four empirical equations (linear, exponential, power, and logistic) and two growth functions (von Bertalanffy and Gompertz) were fitted to the ML-at-age data. The equations are as follows.

Linear function:

$$ L = a + bt, $$
(1)

exponential function:

$$ L = a \times e^{bt} , $$
(2)

power function:

$$ L = a \times t^{b} , $$
(3)

logistic function:

$$ L = \frac{a}{{1 + {\text{Exp}}( - b(t - c))}}, $$
(4)

von Bertalanffy function:

$$ L = a \times (1 - {\text{Exp}}( - b(t - c))), $$
(5)

Gompertz function:

$$ L = a \times {\text{Exp}}(b \times (1 - {\text{Exp}}( - ct))), $$
(6)

where L is mantle length (mm) at age t; t is age (day); a, b, c, are parameters to be estimated.

The Schnute function [29, 30] was also used.

Schnute function:

$$ L = \left[ {L_{1}^{{g_{2} }} + \left( {L_{2}^{{g_{2} }} - L_{1}^{{g_{2} }} } \right) \times \left( {\frac{{1 - {\text{Exp}}( - g_{1} \times (t - t_{1} )}}{{1 - {\text{Exp}}( - g_{1} \times (t_{2} - t_{1} )}}} \right)} \right]^{{(1/g_{2} )}} , $$
(7)

where t 1 and t 2 are the minimum and maximum ages of squid sampled, respectively. L 1 and L 2 are the estimated size at the minimum and maximum observed ages, respectively. Parameters g 1 and g 2 are growth rates to be estimated.

The parameters of the models were estimated by the least squares method (linear models) and non-linear least squares using the Gauss–Newton algorithm (non-linear models) incorporated in software R (Version 3.1.1; www.r-project.org).

The best growth function was identified using the Akaike information criterion (AIC) [31]:

$$ {\text{AIC}} = 2{\text{LL}} + 2P, $$

where LL is the negative log-likelihood and P the number of parameters in each function.

Maturity ogives were estimated by fitting the logistic curve: Y = 1/(1 + Exp(−a(X − b))) to the proportion of mature squid (Y) by age or ML interval (X); where a is the slope of the curve and b is the age (t 50 %) and ML (ML50 %) at which 50 % of females or males were mature [32]. Parameters a and b were calculated by non-linear least squares using the Gauss–Newton algorithm incorporated in software R (Version 3.1.1).

Principal component analysis (PCA) was performed to extract the variation in six life-history traits (BW, ML, number of increments, statolith radius, average growth rate and GSI) between monthly groups (micro-cohorts) of the squid. The score of the first two PCs of each group was applied to analyze the similarity of growth parameters between groups by cluster analysis with Euclidean distance and unweighted pair-group average method.

Results

Of the 341 S. lessoniana specimens collected, 271 statoliths (79.5 %) were appropriately prepared for reading (Table 1), which consisted of 142 females and 129 males. ML ranged from 88 to 355 mm and from 85 to 401 mm for females and males, respectively. No spent individual was found of either sex. The smallest individuals were collected in August for both sexes and the largest ones were collected in March (male) and April (female, Fig. 2). The estimated coefficients presented an allometric relationship between ML and BW of S. lessoniana. The slopes of the ML-BW relationship differed significantly between the two sexes (ANCOVA, F 1, 267 = 4.94, P = 0.027); females are heavier than males at the same ML (females: BW = 0.0003ML2.70, R 2 = 0.995; males: BW = 0.0003ML2.65, R 2 = 0.994).

Table 1 Summary of the collection of statoliths of Sepioteuthis lessoniana in the waters off northern Taiwan between April 2009 and March 2010 (F: females, M: males)
Full size table
Fig. 2
figure 2

Monthly variation in mantle length, shown by box-whisker plot, of Sepioteuthis lessoniana in the waters off northern Taiwan. (closed circles are outliers)

Full size image

Age and hatching dates

Of the 271 prepared squid, estimated ages ranged from 77 to 216 days. The youngest female, immature at 77 days and 100 mm ML was collected in August; and the youngest male, immature at 86 days and 85 mm ML was collected in August. The oldest female, mature at 216 days and 355 mm ML was collected in April; and the oldest male, mature at 209 days and 345 mm ML was collected in April. Squids collected from April 2009 to March 2010 had hatched almost year-round, except in January–February (Fig. 3). The hatching month peaked in August and September, while a minor peak in May was noted.

Fig. 3
figure 3

Frequency distribution of back-calculated hatching dates of Sepioteuthis lessoniana in the waters off northern Taiwan

Full size image

Growth pattern

All parameters of growth functions fitted to ML-at-age data of the squid are tabulated in Table 2. The parameters for both sexes failed to converge when fitted to a von Bertanlanffy growth function, although the parameters can be estimated. The age-at-ML data was best described by Schnute function (followed by logistic function) and linear equation (followed by logistic function) based on the AIC values (Table 2). The scatter-plot of ML-at-age data for females showed an asymptotic pattern (Fig. 4a), while that for males showed a high variation after 130 increments (Fig. 4b).

Table 2 Parameters and Akaike’s information criterion (AIC) of fitted growth functions of the mantle length-at-age data of Sepioteuthis lessoniana in the waters off northern Taiwan
Full size table
Fig. 4
figure 4

Mantle length-at-age data and best described growth functions (solid lines) of Sepioteuthis lessoniana in the waters off northern Taiwan. (a females, b males)

Full size image

Age- and ML-at-maturity

Age- and ML-at-maturity for both sexes were estimated by fitting logistic curves to the proportion of mature squid by age (in days) or size class. Maturity ogives by age group explained 97.7 and 93.4 % of the variance in maturity for females and males, respectively (Fig. 5a). Females matured at 153 days, while males matured ca. 10 days earlier (at 143 days) than females. Maturity ogives by size class explained 91.4 and 96.5 % of the variance in maturity for females and males, respectively (Fig. 5b). ML-at-maturity was estimated to be 244 and 232 mm for females and males, respectively.

Fig. 5
figure 5

Proportion of mature individuals by a age, and b mantle length class and estimated logistic curves of Sepioteuthis lessoniana in the waters off northern Taiwan

Full size image

Growth and maturation for seasonal cohorts

Two seasonal cohorts were distinguished by natural break and information from local fishermen in distribution of hatching month for S. lessoniana (Fig. 3): the spring cohort (SC) hatched from March to July, and the autumn cohort (AC) hatched from August to December. ML-at-age data of AC females were best described by a power equation and those data of others (AC males and both sexes for SC) were best described by linear equations (Fig. 6).

Fig. 6
figure 6

Mantle length-at-age data and linear functions of females (left) and males (right) by autumn cohort (AC, a and b) and spring cohort (SC, c and d) of Sepioteuthis lessoniana in the waters off northern Taiwan

Full size image

Maturity ogives of females by age group explained 95.9 and 96.8 % of the variance in maturity for females of AC and SC, respectively (Fig. 7a). Age-at-maturity was estimated to be 150 and 154 days for AC and SC, respectively. Maturity ogives of males by age group explained 76.5 and 98.2 % of the variance in maturity for AC and SC, respectively (Fig. 7b). Age-at-maturity was estimated to be 138 and 141 days for AC and SC, respectively. Maturity ogives of females by size group explained 94.0 and 53.0 % of the variance in maturity for females of AC and SC, respectively (Fig. 7c). ML-at-maturity was estimated to be 248 and 237 mm for AC and SC, respectively. Maturity ogives of males by size group explained 86.7 and 90.6 % of the variance in maturity for females of AC and SC, respectively (Fig. 7d). ML-at-maturity was estimated to be 231 and 240 mm for AC and SC, respectively.

Fig. 7
figure 7

Proportion of mature individuals by age (a–b), and mantle length class (c–d) and estimated logistic curves for autumn (AC) and spring cohorts (SC) of Sepioteuthis lessoniana in the waters off northern Taiwan

Full size image

In order to extract the variation between monthly groups (micro-cohorts) of the squid, six life-history traits (BW, ML, number of increments, statolith radius, average growth rate and GSI) were selected to perform principal component analysis (PCA). The first two factors (PC1 and PC2) explained 85.9 and 87.6 % of the total variance in life-history traits for females and males, respectively (Table 3). The first factor (PC1) was positively correlated with six variables (BW, ML, NI, SR, GSI and average growth rate, GR) for females and males, while the second factor (PC2) was positively correlated with GR and GSI for females and males, respectively. The monthly groups from the same season for both sexes of the squids showed similar distribution patterns in the bi-plot composed of PC1 and PC2, AC in the positive direction along PC1, whereas SC was in the negative direction along PC1 (Fig. 8). Monthly groups of the squids showed two extreme conditions of growth parameters: smaller size and slower growth rate for groups hatched in March–May (negative PC1), whereas larger size and faster growth rate for groups hatched in September–October (positive PC1). Nevertheless, squids hatched in June–August and November showed a transitional pattern in growth parameters between these two extreme conditions. Dendrograms derived from cluster analysis based on PCA scores illustrated the similarity in growth parameters between monthly groups of the squids (Fig. 9). In general, squids hatched in the two peak seasons (March–May vs September–October) were of high similarity in growth parameters, while those hatched in transitional months were of high variation in growth parameters. Clusters of monthly groups of the squids could correspond to seasonal cohorts in females, whereas a high similarity of growth parameters was found for squids hatched in transitional months in males.

Table 3 Loading, eigenvalues and percent variance for two principal components (PC) from principal component analysis for the Sepioteuthis lessoniana in the waters off northern Taiwan based on six biological measurements
Full size table
Fig. 8
figure 8

Scores (in average ± standard error by month) and loadings in two principal components from principal component analysis for spring cohort (open ellipses) and autumn cohort (shaded ellipses) of Sepioteuthis lessoniana in the waters off northern Taiwan based on six biological measurements. (BW body weight, ML mantle length, NI number of increments, SR statolith radius, GSI gonadosomatic index, GR average growth rate)

Full size image
Fig. 9
figure 9

Similarity of the monthly groups of Sepioteuthis lessoniana in the waters off northern Taiwan. (a females, b males)

Full size image

Discussion

Sepioteuthis lessoniana off northern Taiwan hatched almost year-round. The ML-at-age data were best described by Schnute and linear functions for females and males, respectively. The females mature later and at a larger size than the males, although the males attained a larger size-at-age than the females. Growth and maturation parameters of two seasonal cohorts (SC and AC) were further analyzed: AC tended to grow faster while maturing at a larger size than SC. Nevertheless, the life-history traits of the seasonal cohorts were complicated and could be inferred by progressive changes in the life-history traits of the squids hatching in the transitional months, i.e. months between two peak seasons. These results provide essential information for S. lessoniana populations off northern Taiwan.

Age and hatching dates

Age and growth of S. lessoniana has been studied for populations from tropical to sub-tropical waters [17, 20, 21]. Geographical differences in estimated maximum life span of S. lessoniana have been shown in previous studies, with estimated maximum age at 124–136 days for squids in the Philippines and Thailand waters, while maximum age was 153–224 days for squids in tropical and subtropical Australian waters [17, 21]. Thus, a short life span (less than 200 days) was suggested for S. lessoniana [12]. In the present study, estimated ages for mature squid ranged from 113 to 216 days, which covered the maximum ages of squids from tropical and subtropical waters. The waters off northern Taiwan are at a latitude of 25°10′N, approximately the same latitude as Hervey Bay in Australia (subtropical waters [17]), although in the opposite hemisphere. The monthly mean sea surface temperature (SST) off northern Taiwan ranged from 19.1 to 28.6 °C between 2008 and 2010 which showed a wider range than the monthly mean temperature in Hervey Bay (from 23 to 27.2 °C in 1998 [17]). The great variation of monthly SST off northern Taiwan may result in a wider range of maximum ages of S. lessoniana.

Sepioteuthis lessoniana in the waters off northern Taiwan hatched almost year-round, excluding January and February. A similar hatching pattern was observed for S. lessoniana populations in Palk Bay in the Indian Ocean [33], whereas specific spawning seasons were suggested for populations in Zanzibar (from July to November [23] ) and in Japanese waters (from April to October [22] ). Year-round hatching of intra-annual cohorts or “micro-cohorts” results in continuous recruitment for short-lived species such as squids [34, 35]. The micro-cohort structuring of squids has been evident in many species, particularly those under commercial exploitation, e.g. Dosidicus gigas in the Gulf of California [36], Illex argentinus in the Southwest Atlantic [37], I. coindetii in the Strait of Sicily [38], and Loligo vulgaris in Galician [39] and Portuguese [40] waters. The composite intra-annual structures and plasticity of life-history parameters of squids may advance adaption and sustainability of the population that are sensitive to changes in environmental conditions and are under fishery exploitation [18, 24, 41, 42].

Growth pattern of S. lessoniana

Most squids are generally short-lived (usually 1 year) and have a semelparous life history, i.e. adults die after spawning [28]. The lifetime energy budget of squids seems to have evolved through physiological progenesis, i.e. the allocation of their energy resources among growth components is an essential characteristic of the early life of iteroparous forms [43]. Theoretically, the whole ontogenesis of any species could be subdivided into several periods of growth (‘stanzas’, [44]), which usually correspond with major ontogenetic events like hatching, sexual maturation, or shifts of habitats [44]. Growth within each stanza of an individual may follow a sigmoid curve, or so-called Sachs cycle [44]. The lower part (early phase) of the growth trajectory may be described by exponential functions, whereas the upper part (final phase) can be described by one of the asymptotic functions [30, 44]. Squids have been identified as having either two (embryonic and post-embryonic) or three growth stanzas (embryonic, paralarval and juvenile-adult) during their ontogenesis [30].

In the present study, the size-at-age data of the entire squid samples (with mixed cohorts) were best described by Schnute and linear functions for females and males, respectively. However, the size-at-age data of seasonal cohorts of the squids were best described by linear or power equations. A power function was suggested for populations of S. lessoniana occurring in tropical and subtropical waters of Thailand and Australia [17]. The asymptotic growth trajectory for the entire squid samples might be a composite pattern from mixing several micro-cohorts which experienced a fast and short life history [17, 45].

Maturation of S. lessoniana

The age-at-maturity of S. lessoniana off northern Taiwan was at ca. 153 and 143 days for females and males, respectively, whereas ML-at-maturity was 244 and 232 mm for females and males, respectively (Fig. 5). A similar ML-at-age was found in Japanese waters (200–250 mm for females [22]), while a smaller size-at-maturity was found in Zanzibar waters (ca. 162 mm for females [23]). It is evident that populations of S. lessoniana in tropical waters are smaller-sized and mature earlier compared to populations in subtropical waters [17]. Squids dwelling in cooler waters could grow more slowly, mature later and have a longer life span than their counterparts in warmer waters, such as Zanzibar.

A higher variability in age-at-maturity than in size-at-maturity of S. lessoniana off northern Taiwan was found for males, while that was not the case for females (Fig. 5). Several studies have found that maturation of squid is more strongly influenced by size rather than age [20, 46]. Maturation of S. lessoniana has been shown to be a size-related process, where the weight of reproductive organs is highly correlated to body weight, rather than with age [18]. A high variation in age-at-maturation of S. lessoniana off northern Taiwan suggested that the population might be composed of multiple inter-annual cohorts experiencing different environmental conditions and with distinct growth rates [18, 40], and hence varying age-at-maturity.

Seasonal variation in growth and maturation

Growth of squids might be influenced by water temperature [47, 48], as well as food availability [16]. Ambient water temperature has been shown to have a positive correlation with growth rates, in particular during early life stages, in a range of squid species, e.g. Loligo pealei [49]; Lolliguncula brevis [46]; Loligo gahi [50]; Sepioteuhtis lessoniana [15]. The differences in growth patterns of squid may reflect differences in life-history traits for populations dwelling in different environmental conditions [18, 51]. Two growth strategies for populations of S. lessoniana from two extremes of environmental conditions have been proposed [17]: the ‘hot’ strategy with a small body size and a short life span for squid in equatorial waters (Thailand), and the ‘cool’ strategy with a larger body size and a longer life span for squid in subtropical waters (Hervey Bay). The life-history traits of S. lessoniana in Townsville, which is located around 19°13′S in the tropical waters between Thailand and Hervey Bay, showed an alternation between two growth strategies depending on season of hatching [17]. This may explain the differences in growth pattern of squid in this study to a degree.

Temporal variability in growth parameters of squids has been shown, with a faster growth rate and higher ML-at-age for squids hatching in the warmer season than for those hatching in the colder season [50, 52]. Nevertheless, environmental conditions may affect growth patterns of squids through the life cycle. The warm cohort of Loligo vulgaris, hatching in warmer waters while subadults and adults live in colder waters, has higher ML-at-age, a longer life span and matures at a smaller size and earlier age than the cold cohort, hatching in colder waters while subadults and adults live in warmer waters [40, 51]. In this study, the squids hatching in warmer waters (from August to October) have higher growth rates and larger sizes than those hatching in colder waters (from March to May, Fig. 8). This suggests that populations of S. lessoniana off northern Taiwan may show similar growth strategies to the seasonal cohorts of L. vulgaris in Portuguese waters. However, little difference in growth and maturation parameters between seasonal cohorts of S. lessoniana was found (Figs. 6, 7). The differences in life-history traits between seasonal cohorts could be inferred by progressive changes in the life-history traits of the squids hatching in the transitional months (June–July for females and June–August for males, Fig. 9), when the oceanographic conditions shift from northeasterly to southwesterly monsoon [53, 54]. Nevertheless, it should be noted that a smaller difference between seasonal cohorts might have been induced by the low sample size; thus, further study may be required to examine explicitly the temporal variability in life-history parameters of S. lessoniana off northern Taiwan.