Experiments were conducted at the Fishery Research Laboratory of Kyushu University (KU) during 2009–2011 and Hakatajima Field Station of the National Research Institute of Fisheries and Environment of the Inland Sea (NRIFEIS) during 2016–2019 ( Supplementary Table S1 ). Chub mackerel were caught from the wild using purse seine or fixed net in coastal waters off western Japan and reared for 0.5–2.5 years in sea pens at a fish farm in Oita Prefecture, western Japan. A total of 216 specimens were transferred to KU or NRIFEIS and moved into 3–50 kL concrete tanks circulated with seawater. Fish were assimilated and reared under natural photoperiod (10–14 h) and temperature (11–17°C) for 1–6 months before the onset of experiment. For experiments conducted at KU, all specimens (n = 83) were from fish farms, whereas for those at NRIFEIS, in addition to specimens from fish farms (n = 133), a total of 211 artificial seedlings at age 1–3 years were used to examine the effects of maternal age on egg and larval traits. The water temperature in the tank was monitored every hour using a data logger (Tidbit V2; Onset, MA, USA). Fish were daily fed with 1–1.5% of their body weight of commercial dry pellets (Marubeni Nisshin Feed Co., Ltd., Tokyo, Japan; Otohime EP 6: protein 52%, oil 11%, ash 18%, fiber 3%).

Spawning was induced based on a sustained-release GnRH delivery system protocol adopted from Shiraishi et al. (2008) and Nyuji et al. (2011; 2012). A total of 427 specimens were anesthetized with 2-phenoxyethanol (300 ppm). Of these, females with late vitellogenic oocytes (600–650 μm diameter), which were determined using measurement of oocyte diameter and clearing solution method (Yoneda et al., 2013), were selected after ovarian biopsy using a plastic catheter tube (Internal diameter: 2 mm). Males that oozed milt under gentle abdominal pressure were selected. After selecting the required number of females and males, females were intramuscularly injected with 400 μg/kg body weight GnRH agonist (D-Ala6, des-Gly10)-LHRH ethylamide (Sigma-Aldrich, St. Louis, USA) combined with coconut oil. The first spawning was observed 34–36 h post-injection, and subsequent daily spawning occurred between 22.00 and 24.00 h. In the following experimental system, daily spawning of chub mackerel was observed for 12–28 d after the injection. A total of 83 specimens (age 1 year, n = 20; age 2 years, n = 33; age 3 years, n = 30) were used for examination of the egg size–temperature relationship under ambient temperature while a total of 344 specimens (age 1 year, n = 139; age 2 years, n = 75; age 3 years, n = 130) were used for examination of the egg size–number of spawning events relationship at a constant temperature (19°C).

Adult specimens randomly selected (n = 312) were measured to within 0.1 mm of fork length (FL) and within 0.1 g of body weight (BW) before and after the experiments. Of these, a total of 209 specimens were measured to within 0.1g gonadal weight (GW) after the experiments. Fulton’s condition factor (K) is the morphometric condition factor of fish species representing the mass of an individual relative to its body length (Ricker, 1975). K was calculated using the following equation:

where SW (somatic weight, g) was calculated as BW – GW. From the R package of lme4 (Bates et al., 2015), linear mixed effects models (LMEM) with FL, BW, and condition factor as fixed effects and experimental trials as random effect indicated that FL and BW of 1-year-old specimens were significantly smaller than those of 2- and 3-year-old specimens ( Supplementary Figure S1 , P< 0.01), but no significant difference in CF among the three age groups was found (P = 0.08).

To examine the effect of temperature and spawning events on egg size, 1-, 2-, and 3-year-old adult females were reared at ambient water temperature of 19°C (range: 18.5–19.5°C) after the GnRHa treatment. During the experiments, fish were daily fed with 1–1.5% of their body weight of commercial dry pellets. Egg diameter was measured every spawning day using Image-Pro Plus ver. 7.01 (Media Cybernetics, Inc., Bethesda, MD, USA) after fixing in 10% neutral sea water formaldehyde. Some eggs were frozen for analysis of the energy content. Egg size was determined using the following equation:

where r is egg radius (mm). To ensure the effects on egg size, the experiments were conducted twice or thrice in different years of spawning, including specimens from the first to third spawning seasons. A total of 2790 eggs (age 1 year, n = 600, age 2 years, n = 900, age 3 years, n = 1290) were used for analysis of the egg size-temperature relationship while a total of 5040 eggs (age 1 year, n = 2010, age 2 years, n = 1110, age 3 years, n = 1920) were used for analysis of the egg size-number of spawning events relationships.

To determine the energy content of the eggs to be spawned, the concentrations of C and N per egg were determined in the eggs spawned between May and July 2019. Eggs were thawed at room temperature (18–22°C) and rinsed with distilled water; thereafter, a total of 200 eggs were pooled for each egg sample to ensure adequate numbers for the analysis and dried at 60°C in a micro-tube for 3–4 days. A total of 138 egg samples (age 1 year, n = 66; age 3 years, n = 72) were used for this analysis. The dried samples were ground to a fine powder and placed in tin capsules. Lipids were retained to assess the total energy in the eggs. The total C and N were analyzed using an elemental analyzer (Flash EA1112, Thermo Fisher Scientific Co., Ltd. MA, USA). The concentration of C or N in an egg was estimated by dividing the total amount of C or N by the total number of eggs used.

To examine the effect of maternal age and temperature on larval size, approximately 100 eggs to be spawned from 1- and 3-year-old adult females at 18, 20, and 22°C were placed in 30 L tank at identical spawning temperatures in duplicates. The timing of larval hatch of this species varied at different temperature regimes (Hunter and Kimbrell, 1980). Larvae were collected on the day of first feeding (age: 3 days at 20 and 22°C and 4 days at 18°C) and stored in 10% neutral sea water formaldehyde. The timing of first feeding was determined as that after >50% of larvae started feeding. Standard length (SL), eye diameter, body depth at anus, head height, and pectoral body depth ( Supplementary Figure S2 ) were measured using Image-Pro Plus.To determine the effect of maternal age and temperature on starvation tolerance in early life stage, approximately 800 eggs spawned from 1- and 3-year-old females at constant temperature (19°C) were placed in 30 L tank at 20°C until hatching. A total of 30–80 hatchlings were transferred into 1 L glass beakers at three temperature treatments (18, 20, and 22°C) in duplicates. The number of dead larvae in each glass beaker was counted at intervals of 7–14 h, and the experiment was terminated when all the larvae were dead. To determine the effect on starvation tolerance, this experiment was performed twice in different years of spawning ( Supplementary Table S1 ).

To examine the effects of maternal age and temperature on early growth rates, approximately 5000 eggs spawned from 1- and 3-year-old adult females at constant temperature (19°C) were transferred to 1 kL tanks with three temperature treatments in duplicates (target temperature: 17, 20, and 23°C). A 14/10-h light/dark cycle was maintained. The water temperature in the tanks was monitored hourly using a data logger. After the eyes of the larvae were pigmented, S-type rotifer (Brachionus plicatilis) complexes cultured with highly unsaturated fatty acid-enriched Chlorella vulgaris (Super Chlorella V12; Chlorella Industry Co., Ltd., Fukuoka, Japan) were provided at 20 rotifers per mL in 1 kL tanks from age 2 or 3 days. Rotifer density was monitored daily and maintained by adding new rotifers when necessary. Kaneko et al. (2019) suggested that the thermal environment experienced during the first 10 days after spawning may significantly influence the recruitment success of chub mackerel in the western North Pacific. Thus, the experiment was terminated at 10 d of age. After anesthetizing at first feeding with 2-phenoxyethanol (300 mg/l), the larvae were preserved in 10% neutral sea water formaldehyde. Daily growth rate (G1) was calculated as follows:

The mean SL at first feeding was calculated from the specimens at 3 or 4 days of age in each tank.

Approximately 2500 and 5000 eggs spawned from 2- and 3-year-old females were transferred into 0.5 and 1 kL tanks, respectively, at different temperature regimes ranging from 17 to 23°C, in which a total of 11 trials were conducted. The experiment was terminated when the larvae were 25-day-old because it was the approximate duration associated with recruitment success of chub mackerel. Approximately 20-day-old larvae of 20 mm FL appear to start migrating to the Kuroshio–Oyahshio transition area (Kamimura et al., 2015); individuals with high initial growth until 27 days of age appear to proactively enter the cooler nutrient-rich area for further rapid growth (Higuchi et al., 2019). The methods of rearing fish in the early stages are mentioned above. From age 14–25 days Artemia sp. nauplii enriched with highly unsaturated fatty acids (Super Capsule Powder, Chlorella Industry Co., Ltd.) were provided at 0.2–1.5 Artemia mL⁻¹ day⁻¹. Specimens were collected at the first feeding (age 3 or 4 days) and at the end of the experiment (age 25 days) to estimate the daily growth rate of individuals and determine developmental stages. Daily growth rate (G2) was calculated as follows:

The mean SL at the first feeding was calculated from the specimens in each tank. Specimens at 25 days of age were classified into three stages: larva (SL< 12 mm), juvenile (SL: 12–30 mm), and post-juvenile (SL ≤ 30 mm), according to Watanabe (1970) and Taga et al. (2019); in the post-juvenile stage, individuals usually start swimming in schools. Data on the daily growth rates of chub mackerel collected from the Kuroshio Extension to the Kuroshio–Oyashio transition areas during 2002–2011 (Kamimura et al., 2015) were used to estimate the experienced temperature until the age of 25 days. These specimens were considered the main recruit fish in each sampling year, as their growth rates in the April cohort, most of their specimens, explained significant variations in the abundance of recruitment and recruitment per spawning stock biomass.

The effects of temperature, number of eggs spawned, and maternal age on egg size were analyzed using the LMEM with restricted maximum likelihood estimation. During the spawning cycle, vitellogenesis is usually completed on the previous day of spawning, and subsequently, oocyte maturation and ovulation proceed on the day of spawning (Nyuji et al., 2011; Nyuji et al., 2012). Thus, to examine the effect of temperature on egg size, the average temperatures on the day before and on the day of spawning were used. The models included temperature or the number of eggs spawned and maternal age as fixed effects and experimental trials as random effect.

Maternal age was initially a 3-level factor representing year of age (age 1, 2 or 3 years). The significance of the fixed effects was tested using the R package of lmerTest, Satterthwaite approximation (Kuznetsova et al., 2017). The model of each factor with and without natural logarithmic transformation was selected based on the minimum conditional Akaike information criterion using the R package of cAIC (Vaida and Blanchard, 2005) in addition to the selection of grouping of maternal age classes. To determine whether egg size was more strongly influenced by temperature or the number of spawning events, chi-square test was performed to analyze deviance. The amounts of C and N per egg were compared between eggs spawned from 1- and 3-year-old females using LMEM, including sampling days as random effect.

Maternal age was a 2-level factor representing year of age (age 1 or 3 years). The effects of maternal age and temperature on the SL of first-feeding larvae were examined using LMEM with maternal age (2-level factor as abovementioned) and temperature treatment (3-level factor; 18, 20 or 22°C treatment) as fixed effects and tank as a random effect.

To examine the effect of maternal age on the starvation tolerance of hatched larval fish, a generalized LMEM incorporating binominal response distribution with a logit link function was performed using the R package of glmmML (Broström and Holmberg, 2011). The full models combined maternal age (2-level factor as abovementioned), time after hatching and temperature treatment (3-level factor as abovementioned) as fixed effects and the tank as random effect.

The effects of maternal age on the relationship between body parts and SL of first-feeding larvae were examined using LMEM with SL and maternal age (2-level factor as abovementioned) as fixed effects and tank and temperature treatment as random effects.

The effects of maternal age and temperature on SL-days after hatching relationship andG1were examined using LMEM, with maternal age (2-level factor as abovementioned) and temperature treatment (3-level factor; low, middle or high temperature treatment) as fixed effects and tank as a random effect.

The model of each factor, with and without natural logarithmic transformation, was selected based on the minimum conditional Akaike information criterion. All model fittings were conducted using backward elimination.

The effects of temperature and ontogenetic developmental stage on G2 were examined using LMEM. The full model combined temperature and developmental stage (3-level factor; larval, juvenile or post juvenile stage) as fixed effects and trial as random effect.

Model selection was performed as described above. Statistical analyses were performed using R version 4.0.3 (R Core Team, 2021) and GraphPad Prism 8 (GraphPad Software, LLC, CA, USA).

A negative correlation was observed between temperature and size of eggs of adult females of all age groups reared in ambient water temperature ( Table 1 and Figure 3A ). However, there was a significant difference in the regressions of the egg size–temperature relationships between eggs of 1-year-old and 2- and 3-year-old females. Size of eggs spawned from 1-year-old females was approximately 79% of that from females aged 2–3 years at the tested temperatures. At ambient temperature, significant effects of the number of spawning events and maternal age on egg size was observed; however, egg size was more influenced by temperature (slope = -0.026) than the number of spawning events (slope = -0.004; χ² = 31.28, P< 0.001). At constant water temperature (19°C), there was a negative correlation between egg size and the number of spawning events ( Figure 3B ), and the egg sizes spawned by 1-year-old females were significantly smaller than those of 2–3-year-old. During the spawning events observed, the estimated slope (-0.002) of the regression model at a constant temperature was approximately 8.2% of the relevant model (-0.026) in water at ambient temperature.

There were significant differences in the amounts of C and N per egg between 1-year-old female eggs and 3-year-old female eggs under constant temperature conditions with respect to differences in egg size, but no significant difference in the C/N ratios per egg was observed between the two age classes ( Table 2 and Figure 4 ). LMEM showed that the estimated egg size, amount of C in eggs, and amount of N in eggs of 1-year-old females (0.44 mm³, 15.34 µg, and 3.12 µg) was approximately 78, 76, and 75% of those of 3-year-old females (0.57mm³, 20.27 µg, and 4.16 µg), respectively. LMEM showed that the estimated C/N ratio per egg in the two female age classes was 4.89.

There were clear effects of maternal age and spawning temperature on the SL of the larvae at the onset of feeding ( Table 3 ). Temperature negatively affected the SL of first-feeding larvae, but the SL of larvae derived from 3-year-old females was larger at the tested temperatures than that of 1-year-old females. The mouth width, head height, eye diameter, and body depth at anus of the first-feeding larvae were positively correlated with SL, but there was no significant correlation between pectoral body depth and SL ( Table 4 and Figure 5 ). The effects of maternal age on the relationship of SL with body depth at anus and pectoral body depth were significant, indicating that the body sizes of first-feeding larvae derived from 3-year-old females were larger than those from 1-year-old females.

There were clear effects of maternal age and temperature on the survival rate of larvae under starvation from hatching ( Table 5 and Figure 6 ). Temperature negatively affected the time to survival, but the time to survival of 50% larvae derived from 3-year-old females was significantly longer at the tested temperatures than those from 1-year-old females.

During the period of 0–10 days after hatching, the SL of larvae derived from 3-year-old females was significantly higher than those from 1-year-old females at three different temperature treatments ( Table 6 and Figure S3 ). The estimated SL of larvae at age 10 days derived from 1-year-old females were approximately 85% of those from 3-year-old females in low, middle, and high temperature treatments. The effects of maternal age and temperature on growth rate were obvious ( Table 6 and Figure 7 ). LMEM indicated that the estimated growth rates of larvae derived from 3-year-old females in low, middle, and high temperature treatments were 55%, 31%, and 20% higher than those from 1-year-old females, respectively.

The effect of temperature on ontogenetic development at 25 days of age was obvious ( Figure 8A ). In water at 17.0°C, all 25-day-old specimens were in the larval stage, while juvenile fish were observed in water at 17.2°C. Specimens in post-juvenile stage appeared in water at 18.0°C, and all stages were found in water above 22°C. The developmental stage and temperature significantly affected the daily growth rate in the 25-day-old specimens ( Table 7 and Figure 8B ). The range of estimated daily growth rate at larval, juvenile, and post-juvenile stages were 0.18–0.24 mm/day (17.0–17.2°C), 0.47–1.29 mm/day (17.2–20.3°C), and 1.08–2.47 mm/day (18.0–23.2°C), respectively. In 25-day-old wild fish, the estimated developmental stage was juvenile (0.63) or post-juvenile (0.37), and none of the fish were at the larval stage. The 5th and 95th percentiles of growth rate of 25-day-old wild fish (n = 520) were 0.79 and 1.71 mm/day, respectively. The estimated range of temperature at the juvenile stage, covering the range of growth rates of wild fish, was from 18.4 to 20.3°C, while that at post-juvenile stage was from 18.0 to 20.3°C.

This study showed that the egg size of chub mackerel was influenced by both maternal age and temperature. Temperature showed commonalities in 1–3-year-old females, with a decrease in egg size in water with increasing temperature. However, there was also a negative correlation between egg size and the number of eggs spawned by females maintained at a constant temperature. A decline in egg sizes of individual females as spawning progressed has been observed in some multiple-batch-spawning capital breeders, referring to the situation in which reproduction is supported using energy stored in advance of the spawning season (McEvoy and McEvoy, 1991; Kjesbu et al., 1996; Trippel, 1998; Rideout et al., 2005), probably because of the adjustment of depleted energy by the end of the spawning season (McBride et al., 2015). Chub mackerel appears to be a capital breeder, as the lipid content in its muscle, with peaks in autumn and winter, decreases with gonadal development in spring (Yoshimitsu et al., 2018). However, the regression slope of egg size–temperature relationship was significantly higher than that of the egg size–number of spawning events at ambient temperature and between ambient and constant temperatures, indicating that temperature could be more influential than the number of eggs spawned. The negative correlation between egg size and temperature may be because of a direct physiological response in which the temperature experienced by a female during vitellogenesis influences the size of egg she produces, as has been found in many fish species (Laptikhovsky, 2006; Yoneda et al., 2014; Barneche et al., 2018b). Meanwhile, the effects of maternal age on the size and nutritional contents of eggs were evident as eggs derived from 2- and 3-year-old females (repeat spawners) were significantly larger and more nutritious than those from 1-year-old females (first-time spawners). This partly agrees with the assumption that larger and older females reproduce disproportionately more than smaller and younger females (Hixon et al., 2014; Barneche et al., 2018a), as there were significant differences in the FL and BW of specimens between the two age classes. In captive Atlantic cod (Gadus morhua) controlled for parental history, the spawning experience of individuals influences egg characteristics more than body size because the egg sizes spawned from second- and/or third-time (repeat) spawners are significantly larger than those from first-time spawners (Kjesbu et al., 1996; Trippel, 1998). In this study, the relationship between egg size and the number of spawnings under constant temperature conditions were monitored from the first to the third spawning season. Although body size might be a co-influential factor for variation in egg size, maternal spawning experience, rather than maternal age, could have significant effects on the size and nutritional content of chub mackerel eggs.

The effects of maternal spawning experience on larval starvation tolerance after hatching and the body size of the first-feeding larvae were obvious. Larvae derived from repeat spawners exhibited higher starvation tolerance, SL, and body heights within a given SL than those derived from first-time spawners. Large body sizes of larvae may be beneficial for swimming ability and prey selection, as SL showed positive correlations with swimming speed and mouth width in chub mackerel larvae (Hunter and Kimbrell, 1980). Moreover, even under ad libitum feeding conditions at a given temperature, larvae derived from repeat spawners grew significantly faster until 10 days of age than those from first-time spawners. These results suggest that the large and nutrient-rich eggs from repeat spawners lead to the enhancement of starvation tolerance and growth potential in the early life stages. Our preliminary observations showed that even within the same SL range, 10-day-old larvae derived from repeat spawners swam significantly faster than those from first-time spawners (Nakamura, personal communication). These results appear to be inconsistent with the observations on haddock (Rideout et al., 2005). Differences in larval morphology induced by differences in egg size spawned over the season persisted for about 5 days after hatching, but the effect of egg size on larval growth was only found at low prey densities and not high prey abundance. In Sebastes spp., which are livebearers with internal fertilization, although there was no effect of maternal age or size on the hatchling size, the larvae derived from older females had larger oil globules than those from younger ones; the larvae from older females survived longer under starvation and grew faster in later life than those from younger ones (Berkeley et al., 2004; Sogard et al., 2008).

The effects of temperature on the ontogenetic development and growth rate of chub mackerel at 25 days of age were obvious, as found in previous studies involving wild and captive chub mackerels (Hunter and Kimbrell, 1980; Kamimura et al., 2015; Taga et al., 2019). Higher temperatures under ad libitum food availability led to more advanced development and faster growth than lower temperatures, probably accompanied by an increase in daily food rations (Hunter and Kimbrell, 1980). Our findings also showed that the transitional periods in developmental stages appeared to occur in a limited range of temperatures; larvae were only found at<18°C, whereas juveniles were observed in waters at 17–20°C. As observed by Hunter and Kimbrell (1980), cannibalism toward slower growing fish was commonly observed after specimens reached the juvenile stage. Further, cannibalism continued in the post-juvenile stage as they began to swim in schools in this study. In a given water temperature, faster growing individuals probably acquired more energy than other individuals, including from cannibalism, in a tank. Thus, the evaluation of the relationship between growth rate and temperature at the larval and juvenile stages presented here may be limited. However, all specimens at age 25 days in water at 17.0°C were at larval stage, suggesting that temperatures below 17°C experienced until 25 days after hatching may be a limiting factor for the transition from larval to juvenile stages (Pörtner et al., 2006; Pörtner and Peck, 2010).

Specimens collected at 25 days of age in the Kuroshio–Oyashio transition area were considered either at the juvenile or post-juvenile stage, and no fish were found at the larval stage. Moreover, the growth rates of 25-day-old wild fish ranging from 0.79 to 1.71 mm/day corresponded to estimated temperature of 18.0–20.3°C from the growth rate–temperature relationships at juvenile and post-juvenile stages. Our estimated temperature of the wild fish almost agrees with those estimated from bioenergetic individual-based model (18.5–20.1°C) applied to chub mackerel in the Northwest Pacific (Guo et al., 2022). However, Guo et al. (2022) suggested that there was no significant effect of temperature on early life growth and prey conditions were the dominant environmental factors controlling growth. It is noted that specimens collected in the Kuroshio–Oyashio transition area had already survived the larval stage and migrated from the coastal waters off eastern Japan (around main spawning grounds, Figure 2 ). In chub mackerel, swimming and feeding abilities rapidly develop around the transition from larval to juvenile stages in association with the precocious development of the digestive system (Kohno et al., 1984; Park et al., 2015). Such ontogenetic shifts could assist growth acceleration after the juvenile stage, at a given temperature, as mentioned above. In wild chub mackerel larvae, a higher growth rate could be accompanied by either higher temperature or both higher temperature and better prey availability (Sassa and Tsukamoto, 2010; Kamimura et al., 2015), but the importance of the prey environment increases as individuals grow in the early stages of life (Taga et al., 2019).

Previous studies on the relationship between early growth rate and recruitment indexes of chub mackerel in the western North Pacific support the assumption of growth-selective mortality (Kamimura et al., 2015; Kaneko et al., 2019; Taga et al., 2019; Takasuka et al., 2020; Guo et al., 2022) as found in Atlantic mackerel Scomber scombrus (Robert et al., 2007); faster growth in the early life stages leads to better recruitment. This is likely to be accompanied by the finding that even within the same range of SL, fast-growing individuals of juvenile chub mackerel showed higher swimming ability (burst speed) than the slow-growing conspecifics, probably owing to differences in morphological traits (Nakamura et al., 2022). Offspring have to migrate from coastal waters off Japan, spawning grounds, to Kuroshio–Oyashio transition areas, nurseries, and feeding grounds for recruitment ( Figure 2 , Yukami et al., 2021). Faster growing juveniles often experience a cooler water temperature, where they can access a highly nutritious diet, resulting in further rapid growth (Higuchi et al., 2019). Thus, thermal and/or prey environments surrounding larvae and juveniles, which are largely influenced by the directions of the Kuroshio and Oyashio currents (Kaneko et al., 2019; Guo et al., 2022), could be key determinants of variation in the growth rate of individuals and year classes. Our findings on the maternal and thermal contributions to offspring traits provide additional insights into the mechanism of reproductive success of chub mackerel in these waters ( Figure 9 ).

The sizes of eggs and larvae at the first feeding, end of endogenous feeding, and ontogenetic development and growth are differently affected by temperature. Eggs and first-feeding larvae become smaller with increasing temperature, whereas ontogenetic development and growth occur earlier and faster in warmer waters. Meanwhile, the estimated temperature experienced by wild fish (main recruits) until the age of 25 days ranged from 18.0 to 20.3°C, which is equivalent to the temperatures around the spawning grounds during April (Kaneko et al., 2019). This appears to be restricted to a range of spawning temperatures (16–22°C) found in the waters of Pacific coast of Japan (Takasuka et al., 2008b). This implies that the restricted temperature range may be the thermal window that ensures both the egg quality of maternal fish and the growth rate (ontogenetic development) of offspring, which is necessary for the recruitment of chub mackerel. On the other hand, the onset timing and duration of spawning appear to differ between repeat spawners and first-time spawners (Watanabe and Yatsu, 2006; Watanabe, 2010); the repeat spawners are likely to start spawning earlier and longer than the first-time spawners. Interestingly, during the 1970s and 1980s, when relatively higher recruitment was observed ( Figure 1 ), peak spawning was concentrated in April, whereas in the 1990s and 2000s, when lower levels of recruitment were observed, peak spawning shifted to May (Watanabe and Yatsu, 2006; Watanabe, 2010; Kanamori et al., 2019). Our findings showed that larvae derived from repeat spawners were larger on the day of first feeding and grew and swam faster than those from first-time spawners even under identical thermal conditions. Such traits found in repeat spawners and their offspring may act as a buffer against environmental variations (Anderson et al., 2008; Hixon et al., 2014), enhancing the probability of the early survival of chub mackerel in the western North Pacific. This suggests that repeat spawners of chub mackerel could contribute more to reproductive success than first-time spawners, which partly explains the higher levels of recruit fish biomass in these waters in the recent years.

In conclusion, our findings highlight the importance of maternal contribution to the viability of chub mackerel offspring. Maternal spawning experience, rather than maternal age, could have profound effects on the size and nutritional content of eggs of chub mackerel. Large and nutrient-rich eggs laid by repeat spawners developed into large first-feeding larvae, accompanied by higher growth and swimming potentials than those from first-time spawners. Such an advantage in the early life stages can promote rapid growth within a certain limit of the thermal environment and may allow for easier migration from spawning grounds to nursery grounds. This is inconsistent with the traditional assumption that females of different sizes or ages produce the same number and quality of offspring per unit of body weight (Hixon et al., 2014). Because changes in the age structure of spawning stock biomass can affect the variability, productivity, stability, and yield of fish populations, considering maternal effects on population dynamics has become more important for sustainable fishery management of exploited marine fish (Shelton et al., 2012; Shelton et al., 2015; Ohlberger et al., 2022). Our findings could assist in incorporating the effect of maternal spawning experience on offspring viability into more accurate population dynamics projections of chub mackerel in the western North Pacific region.