INTRODUCTION
Symbolophorus californiensis (Eigenmann & Eigenmann) is a typical trans-Pacific transitional water myctophid species, occurring from off Japan to North America, and is one of the dominant mesopelagic fish in the transition region, located between the Oyashio and Kuroshio fronts, of the western North Pacific (Figure 1; Bekker, Reference Bekker and Rass1967; Willis et al., Reference Willis, Pearcy and Parin1988; Sassa et al., Reference Sassa, Kawaguchi, Kinoshita and Watanabe2002; Brodeur & Yamamura, Reference Brodeur and Yamamura2005). The maximum body size is approximately 125 mm standard length (SL), and adults occur primarily in the productive northern transition region (NTR) between the Oyashio front and Subarctic Boundary and southern Oyashio region off Honshu in summer (Figure 1; Watanabe et al., Reference Watanabe, Moku, Kawaguchi, Ishimaru and Ohno1999; Sassa et al., Reference Sassa, Kawaguchi, Kinoshita and Watanabe2002; Brodeur & Yamamura, Reference Brodeur and Yamamura2005). Analysis of otolith increments suggest it takes approximately 550 days to reach 100 mm SL, thus they live for at least two years (Takagi et al., Reference Takagi, Yatsu, Moku and Sassa2006). The daytime habitat of S. californiensis adults is mainly in the 300–600 m layer, with migration to the 0–100 m layer at night for nocturnal feeding (Watanabe et al., Reference Watanabe, Moku, Kawaguchi, Ishimaru and Ohno1999; Yatsu et al., Reference Yatsu, Sassa, Moku and Kinoshita2005; Takagi et al., Reference Takagi, Yatsu, Itoh, Moku and Nishida2009). They are prey for larger animals such as neon flying squid, blue shark, swordfish, northern right whale dolphin, Pacific white-sided dolphin, and northern fur seal (Walker & Jones, Reference Walker and Jones1993; Yonezaki et al., Reference Yonezaki, Kiyota, Baba, Koido and Takemura2003; Watanabe et al., Reference Watanabe, Kubodera, Ichii and Kawahara2004, Reference Watanabe, Kubodera, Ichii, Sakai, Moku and Seitou2008; Kubodera et al., Reference Kubodera, Watanabe and Ichii2007; H. Watanabe & T. Kubodera, unpublished data). Thus, they are one of the key components of the pelagic ecosystem of the transition region.
The sampling localities of Symbolophorus californiensis larvae during the three cruises in the transition region of the western North Pacific. The position of the Kuroshio and Oyashio fronts and Subarctic Boundary (broken lines) and water masses are shown. KUR, OYA, and SB represent the Kuroshio and Oyashio fronts and Subarctic Boundary, respectively.
The southern transition region (STR) of the western North Pacific is an important spawning and nursery ground from spring through to early summer not only for commercially important pelagic fish such as Japanese sardine (Sardinops melanostictus), Japanese anchovy (Engraulis japonicus), chub mackerel (Scomber japonicus), and Pacific saury (Cololabis saira) but also for many myctophid fish, including subarctic, transitional, and subtropical species (Odate, Reference Odate1994; Kubota et al., Reference Kubota, Oozeki and Kimura2001; Sassa et al., Reference Sassa, Kawaguchi, Kinoshita and Watanabe2002, Reference Sassa, Kawaguchi, Oozeki, Kubota and Sugisaki2004, Reference Sassa, Kawaguchi and Taki2007a; Moku et al., Reference Moku, Tsuda and Kawaguchi2003; Brodeur & Yamamura, Reference Brodeur and Yamamura2005; Watanabe, Reference Watanabe2007). Symbolophorus californiensis adults undergo a southward spawning migration from their feeding grounds in the NTR to the STR across the Subarctic Boundary (Sassa et al., Reference Sassa, Kawaguchi, Kinoshita and Watanabe2002, Reference Sassa, Kawaguchi, Oozeki, Kubota and Sugisaki2004). Larvae of S. californiensis are relatively abundant in the warm and productive upper 100 m layer of the STR, during both day and night in spring to early summer (Figure 1; Tsukamoto et al., Reference Tsukamoto, Zenitani, Kimura, Watanabe and Oozeki2001; Sassa et al., Reference Sassa, Kawaguchi, Oozeki, Kubota and Sugisaki2004, Reference Sassa, Kawaguchi and Taki2007a). This species attains a larger size during the larval stage (approximately 24 mm SL) than other myctophid fish (mostly less than 10–20 mm SL) in the western North Pacific, and the duration of the larval period has been found to range from approximately 40–55 days (Moser & Ahlstrom, Reference Moser and Ahlstrom1970; Takagi et al., Reference Takagi, Yatsu, Moku and Sassa2006).
Despite the ecological importance of myctophid fish in the ecosystems of the world's oceans, information on the feeding habits during larval stage is limited (Sabatés & Saiz, Reference Sabatés and Saiz2000; Sabatés et al., Reference Sabatés, Bozzano and Vallvey2003; Conley & Hopkins, Reference Conley and Hopkins2004; Rodríguez-Graña et al., Reference Rodríguez-Graña, Castro, Loureiro, González and Calliari2005). Recently, Sassa & Kawaguchi (Reference Sassa and Kawaguchi2004, Reference Sassa and Kawaguchi2005) described the feeding habits of five myctophid fish larvae, i.e. two subtropical species Diaphus garmani and Myctophum asperum and three subarctic species Diaphus theta, Protomyctophum thompsoni and Tarletonbeania taylori, in the transition region, including the STR, of the western North Pacific during spring to summer. However, information on larval feeding habits of the transitional water species of S. californiensis remains sparse in the STR because of the difficulty in sampling sufficient numbers of specimens during the daytime due to net avoidance of large larvae (Sassa & Kawaguchi, Reference Sassa and Kawaguchi2005).
The goal of this study was to determine the feeding habits of S. californiensis larvae. Feeding incidence, dietary composition, prey size, trophic niche breadth, and number of prey per larvae were examined.
MATERIALS AND METHODS
Sample collection
Larvae were collected during three cruises in the southern transition region (STR) of the western North Pacific in 1997 and 1998 (Figure 1; Table 1). The STR is located between the Kuroshio and Oyashio fronts in the area west of approximately 150°E, and between the Kuroshio front and Subarctic Boundary in the area east of 150°E (Figure 1; Sassa et al., Reference Sassa, Kawaguchi, Kinoshita and Watanabe2002, Reference Sassa, Kawaguchi, Oozeki, Kubota and Sugisaki2004). In May–June, the Oyashio and Kuroshio fronts are defined by the 5 and 15°C isotherms at 100 m depth, respectively (Kawai, Reference Kawai and Masuzawa1972; Odate, Reference Odate1994). The Subarctic Boundary, denoted by the vertical 34.0 isohaline, is observed at approximately 40°N across the transition region (Favorite et al., Reference Favorite, Dodimead and Nasu1976), but is not clearly observable to the west of 150°E, where a southward extrusion of the Oyashio occurs (Sassa et al., Reference Sassa, Kawaguchi, Oozeki, Kubota and Sugisaki2004). Based on these definitions and results of hydrographic observations, the sampling stations for S. californiensis larvae were selected. The sea surface temperature (SST) of the sampling stations ranged from 15.9–18.6°C, corresponding with the reported SST in the STR (Sassa et al., Reference Sassa, Kawaguchi, Kinoshita and Watanabe2002, Reference Sassa, Kawaguchi, Oozeki, Kubota and Sugisaki2004). At each station, daytime and nighttime samples were collected. Tows conducted between 1 hour after sunrise and 1 hour before sunset, and between 1 hour after sunset and 1 hour before sunrise, were considered ‘daytime’ and ‘nighttime’ samples, respectively.
The number of Symbolophorus californiensis larvae examined for gut contents and mean body length (±standard deviation).
BL, body length.
Between 23 May and 15 June 1997, a multi-layer closing net with an 80 cm mouth diameter and a 0.5 mm mesh (MTD net; Motoda, Reference Motoda1971) was towed for 30 minutes at depths within the upper 100 m during a cruise of the RV ‘Hakuho-Maru’ (Ocean Research Institute, University of Tokyo). A rectangular frame trawl with a 2.24 × 2.24 m mouth opening and 1.59 mm mesh net (Matsuda–Oozeki–Hu trawl (MOHT); Oozeki et al., Reference Oozeki, Hu, Kubota, Sugisaki and Kimura2004) was obliquely towed down to approximately 75 m for 40–50 minutes during a cruise of the RV ‘Soyo-Maru’ (National Research Institute of Fisheries Science, Japanese Fisheries Agency) from 1–6 June 1997, and a cruise of the ‘Wakataka-Maru’ (Tohoku National Fisheries Research Institute, Japanese Fisheries Agency) from 7–9 June 1998. Larvae of S. californiensis have a broad depth distribution range of 20–100 m, with peak abundance at 20–50 m (Tsukamoto et al., Reference Tsukamoto, Zenitani, Kimura, Watanabe and Oozeki2001), thus the sampling layer approximately covered their distribution range. Although the length of tow was long (approximately 30–50 minutes) for feeding analysis compared to other studies (e.g. Conley & Hopkins, Reference Conley and Hopkins2004), the myctophid fish larvae usually showed high daytime feeding incidence (66–84%) by this towing duration in the STR of the western North Pacific (Sassa & Kawaguchi, Reference Sassa and Kawaguchi2004, Reference Sassa and Kawaguchi2005). Plankton samples were fixed in 10% buffered formalin–seawater at sea.
Fish larvae of the entire size-range, from first-feeding to transforming stages, are difficult to collect using one type of sampling gear owing to net extrusion of smaller larvae and net avoidance by larger larvae (Sassa & Kawaguchi, Reference Sassa and Kawaguchi2005). Therefore, two types of gear with different mesh sizes (0.5 and 1.59 mm) and towing speeds (MTD; 1.5 knots, MOHT; 4.0 knots) were adopted. For small larvae the MTD was used, and for larger larvae the MOHT was used. Given the maximum body depth of S. californiensis larvae (Moser & Ahlstrom, Reference Moser and Ahlstrom1970), net extrusion would occur for larvae <3.9 mm in notochord length, even using a 0.33 mm mesh net. There was no significant difference between day and night catchability of small myctophid larvae <5 mm in body length (Sassa et al., Reference Sassa, Kawaguchi and Taki2007a). On the other hand, larger larvae, especially those >10 mm in standard length, may be able to avoid nets visually during daytime sampling, although details of the extent larvae are able to avoid nets remains unclear (Sassa & Kawaguchi, Reference Sassa and Kawaguchi2005).
This study pooled the larvae of a wide size-range collected over a wide area of the STR of the western North Pacific in 1997 and 1998 to describe their average feeding habits (Figure 2; Table 2). Sassa et al. (Reference Sassa, Kawaguchi, Oozeki, Kubota and Sugisaki2004) showed that average body length of S. californiensis larvae increased from west to east, between 144° and 173°E, in the STR of the western North Pacific, i.e. the larvae disperse eastward from their western reproductive areas as they grow. This study collected smaller larvae west of 150°E and larger ones east of 150°E (Figure 1; Table 1), corresponding with the larval distribution pattern in the field.
Length–frequency distributions of Symbolophorus californiensis larvae examined for the gut content analysis. n, total number of larvae measured.
Diel change in the feeding incidence (%) of Symbolophorus californiensis larvae in the transition region of the western North Pacific. N, number of larvae examined.
BL, body length.
Laboratory analyses
Intact specimens of S. californiensis larvae were sorted for gut content analysis, and 263 individuals were examined (Table 1). Before dissection, body length (BL) was measured to the nearest 0.1 mm. Notochord length (NL) was measured for preflexion larvae and standard length (SL) for flexion and postflexion larvae. Upper jaw length was measured from the tip of the snout to the posterior end of the maxilla. Hereafter, ‘mouth size’ refers to upper jaw length, unless otherwise specified (Sabatés & Saiz, Reference Sabatés and Saiz2000; Conley & Hopkins, Reference Conley and Hopkins2004; Sassa & Kawaguchi, Reference Sassa and Kawaguchi2005).
Larval size (BL) of S. californiensis ranged from 3.7–22.2 mm (mean±standard deviation: 11.1±3.9 mm) (Figure 2). The yolk is absorbed by the time they reach approximately 3.8 mm NL (Moser & Ahlstrom, Reference Moser and Ahlstrom1970), suggesting that first-feeding occurs around this size. The transformation from larvae to juveniles takes place at approximately 24 mm SL; S. californiensis larvae attain the largest size of any species of myctophid larvae in the western North Pacific (Moser & Ahlstrom, Reference Moser and Ahlstrom1970). This study covered most larval stages after yolk-sac absorption, except for transforming and just before transforming larvae of 22.2–24 mm SL. The absence of the transforming stage relates to the difficulty of collection, possibly due to the much lower density and deeper distribution depth of this stage compared to earlier larval stages (Sassa et al., Reference Sassa, Kawaguchi, Hirota and Ishida2007b).
The gut of each larva was dissected from the body and opened lengthwise with fine needles in a glycerin drop to avoid dispersion of the gut contents. Prey items were counted and identified to the lowest possible taxon. The maximum body width of each prey item (hereafter referred to as prey width) that the larvae would have to encompass for ingestion was measured to the nearest 0.01 mm along the maximum cross-section under a microscope fitted with an ocular micrometer (Blaxter, Reference Blaxter1965).
Data analysis
Gut contents of S. californiensis larvae from different localities and years were examined (Figure 1). All larvae that had identifiable prey in their guts were used for the analysis. Feeding incidence was calculated as the percentage of larvae that had gut contents out of the total number of larvae examined for the daytime and nighttime samples.
Prey items of S. californiensis were evaluated using the percentage of each item out of the total number of diet items examined (%N) and per cent frequency of occurrence of each prey item (%F) in all larvae that had gut contents. The product of these two factors, referred to as %N × %F, was used as an index of the relative importance (IRI) of each prey item. To readily allow comparison among prey items, the IRI was then standardized to %IRI for each prey item i using the following equation (Cortés, Reference Cortés1997):
where n is the total number of prey categories considered at a given taxonomic level.
The larvae of S. californiensis were separated into four size-classes to assess ontogenetic changes in feeding incidence and prey composition with growth, i.e. 3.7–7.9 mm, 8–11.9 mm, 12–15.9 mm and 16–22.2 mm BL. Notochord flexion takes place between approximately 8.5 and 10 mm BL (Moser & Ahlstrom, Reference Moser and Ahlstrom1970). Thus, larvae in the smallest size-class were at the preflexion stage.
Intraspecific differences in prey size were examined throughout larval development. Pearre's trophic niche breadth (Pearre, Reference Pearre1986) was adopted to analyse the relationship between prey size and predator size. This model uses the standard deviation (SD) of the log10-transformed prey size as a measure of the trophic niche breadth. In this analysis, fish larvae were classified according to the body length at 0.1 mm intervals. Only classes with >1 prey item in the gut were used for further analyses. The average and SD of the log10-transformed prey width was calculated for each available size-class of larval fish. The relationship between body length and the corresponding average and SD of the log10-transformed prey size was examined using linear regression analysis to determine any shifts in prey size and niche breadth with growth.
Regression analysis was used to determine the relationship between the number of prey in the gut and larval body length. The relationship between body length and upper jaw length was fitted using linear, allometric, and logarithmic formulae using the least squares method. The Shannon–Wiener diversity index (H′; Shannon & Weaver, Reference Shannon and Weaver1949) was used as a measure of prey species diversity.
RESULTS
Feeding incidence
The smallest larva with gut contents was 5 mm BL; thus, first-feeding larvae were not represented in the samples. Feeding incidence was more than 50% (53.1–92.3%) in all size-classes of S. californiensis larvae during the daytime, while at night it decreased to <6% (Table 2). Daytime feeding incidence tended to gradually increase ontogenetically, from 56.3% at 3.7–7.9 mm BL to 92.3% in the largest size-class (16.0–22.2 mm BL) (Table 2).
Diet composition and trophic ontogeny
Symbolophorus californiensis larvae of the smallest size-class of 3.7–7.9 mm BL, i.e. preflexion larvae, fed mainly on copepod nauplii and copepod eggs, which accounted for 62.4% and 24.6% of the percentage index of relative importance (%IRI), respectively (Table 3). With growth, the importance of these items decreased rapidly, with low %IRI values of below 0.3% in the larger size-classes of ≥12 mm BL (Table 3). Instead, calanoid copepodites became the most important and contributed 53.4, 90.9, and 76.0% of the %IRI values in the larger size-classes of 8–11.9, 12–15.9, and 16–22.2 mm BL, respectively. Although identification of calanoid copepodites to the genus level was not always possible, due to their advanced stages of digestion, Pseudo/Paracalanus spp. (%IRI, 14.8–85.2%) were the most important prey items for the larvae ≥8 mm BL (Table 3). In addition, Eucalanus spp. and Neocalanus spp. tended to become more common with growth, although the respective %IRI values were relatively low (0.3–5.2% and 1.1–3.2%, respectively). On the other hand, occurrence of cyclopoid and poecilostomatoid copepodites were restricted with the low %IRI values of <5% during the larval stage. The furcilia stage of euphausiids was also found in the gut of larvae ≥8 mm BL, and ranked second (%IRI, 17.4%) in the largest size-class. The diversity of prey items was highest in the size-class of 8–11.9 mm BL (H′= 2.65), and then gradually decreased with larval growth (Table 3).
Prey composition of the four size-classes of Symbolophorus californiensis larvae. %N, percentage of each prey item out of the total number; %F, per cent frequency of occurrence of each prey item among larvae with food in their guts; %IRI, index of relative importance; –, no occurrence; BL, body length.
Prey size and trophic niche breadth
The prey width size-range of S. californiensis larvae were 45–1590 µm. The logarithmic average sizes of prey were significantly correlated with larval length (Figure 3; r2 = 0.482, df = 52, P < 0.05). The rate of prey size shift with growth (i.e. slope of the line for the prey size/larval fish length relationship) was higher for larvae ≤11.9 mm BL (0.076) than for larvae ≥12 mm BL (0.038) (ANCOVA, df = 1, F = 5.190, P < 0.05). A positive relationship was observed between larval size and niche breadth (r2 = 0.122, df = 52, P < 0.05), with average niche breadths of 0.13±0.11 at ≤11.9 mm BL and 0.17±0.14 at ≥12 mm BL (Figure 3).
Prey size–fish body length relationship for Symbolophorus californiensis larvae. Prey size was measured as the maximum body width. Lines are standard deviations.
The relationship between body length (BL in mm) and mouth size (MS in mm) was expressed as follows (number of individuals examined = 89):
Prey size (maximum prey width) as a percentage of mouth size ranged from 3.8–73.5% (average±SD, 15.3±9.6%).
Number of prey items per gut
Only daytime samples were examined, since the larvae fed mostly during the daytime (Table 2). The number of prey per gut ranged from 0–14. In larvae ≤11.9 mm BL, there were no significant differences in the number of prey items per gut among the larvae of various sizes (Kruskal–Wallis test, P > 0.1), with the average of 1.1±0.2 (SD) individuals per gut (Figure 4). After 12 mm BL, the number of prey eaten increased with body length (Figure 4; r2 = 0.112, df = 106, P < 0.05). The average number of prey was 2.5 and 4 individuals per gut in the larvae of 12–16 mm and 16–20 mm BL, respectively.
Mean number of prey per gut of Symbolophorus californiensis larvae in relation to body length (mm). Numbers above the plots are the numbers of larvae examined.
DISCUSSION
Diel feeding periodicity
Symbolophorus californiensis larvae fed mainly during the daytime in the epipelagic layer, suggesting that they are visual feeders. This diurnal feeding pattern agrees closely with results for many other myctophid fish larvae reported in the world oceans (Sabatés & Saiz, Reference Sabatés and Saiz2000; Sabatés et al., Reference Sabatés, Bozzano and Vallvey2003; Conley & Hopkins, Reference Conley and Hopkins2004; Sassa & Kawaguchi, Reference Sassa and Kawaguchi2004, Reference Sassa and Kawaguchi2005; Rodríguez-Graña et al., Reference Rodríguez-Graña, Castro, Loureiro, González and Calliari2005), except for Myctophum selenops and Protomyctophum thompsoni that in large size-class larvae begin feeding at night (Conley & Hopkins, Reference Conley and Hopkins2004; Sassa & Kawaguchi, Reference Sassa and Kawaguchi2005). Although myctophid larvae that have been examined have pure rod-like retinas that would enhance feeding efficiency at depth in dim light environments (Pankhurst, Reference Pankhurst1987; Sabatés et al., Reference Sabatés, Bozzano and Vallvey2003), their vision is not used for nocturnal feeding. Symbolophorus californiensis begin diel vertical migration (DVM) and actively feed in the epipelagic layer at night after transformation from larvae to juveniles (Kawamura & Fujii, Reference Kawamura and Fujii1988; Watanabe et al., Reference Watanabe, Moku, Kawaguchi, Ishimaru and Ohno1999; Sassa et al., Reference Sassa, Kawaguchi, Kinoshita and Watanabe2002; Yatsu et al., Reference Yatsu, Sassa, Moku and Kinoshita2005; Takagi et al., Reference Takagi, Yatsu, Itoh, Moku and Nishida2009). Thus, they change their feeding rhythms from diurnal to nocturnal feeding after the transformation at approximately 24 mm SL, which would correlate with changes in morphology and physiology of the visual system during the development (Evans & Browman, Reference Evans and Browman2004).
Daytime feeding incidence of S. californiensis increased gradually with larval growth, from 56% to 92%. The similar pattern has also been documented for seven other myctophid species in the STR and Mediterranean (Sabatés et al., Reference Sabatés, Bozzano and Vallvey2003; Sassa & Kawaguchi, Reference Sassa and Kawaguchi2004, Reference Sassa and Kawaguchi2005). This result indicates that the capture success rate and energy demands increase with growth. Although the feeding incidence could also relate to the digestibility of prey (Fukami et al., Reference Fukami, Watanabe, Fujita, Yamaoka and Nishijima1999; Conley & Hopkins, Reference Conley and Hopkins2004), this would not be so serious factor in this study, since S. californiensis fed mainly on crustaceans throughout the larval stage.
During the nighttime when S. californiensis larvae stop feeding, the gas bladders of all individuals examined contained a large volume of gas and were notably inflated (Figure 5a). In most cases, the inflated gas bladder constricted the gut (Figure 5b). On the other hand, the larvae collected during daytime had a small inconspicuous gas bladder without gas, i.e. non-inflated. A similar phenomenon has been also observed in many other fish larvae, including myctophids (Hunter & Sanchez, Reference Hunter and Sanchez1976; Hoss & Phonlor, Reference Hoss and Phonlor1984; Leis & Carson-Ewart, Reference Leis and Carson-Ewart2000; Govoni & Hoss, Reference Govoni and Hoss2001; Herring, Reference Herring2002; C. Sassa, unpublished data). Although the physiological mechanisms are unknown, this would allow S. californiensis larvae to reduce swimming activity during non-feeding periods while maintaining their depth in the water column, as suggested for other fish larvae (Hunter & Sanchez, Reference Hunter and Sanchez1976; Hoss & Phonlor, Reference Hoss and Phonlor1984).
(A) Inflated gas bladder of a Symbolophorus californiensis larva (15.2 mm standard length) collected at night; (B) the lateral view of the gut of a S. californiensis larva (20.1 mm SL) collected at night. The gut was constricted by the inflated gas bladder from the dorsal direction (indicated by open arrows). Arrows of S, E, and A show inflated gas bladder, oesophagus and anus, respectively. Scale bar = 1 mm.
Diet composition and trophic ontogeny
In the STR of the western North Pacific, S. californiensis larvae depend mainly on copepods of various developmental stages including eggs, nauplii, and copepodites. Dominance of copepods in the gut of myctophid larvae is reported in many other species in various parts of the oceans (Sabatés & Saiz, Reference Sabatés and Saiz2000; Conley & Hopkins, Reference Conley and Hopkins2004; Sassa & Kawaguchi, Reference Sassa and Kawaguchi2005), although there are species that prey mainly on non-copepod plankton. For example, Diaphus garmani larvae depend mainly on appendicularian houses and Myctophum asperum fed mainly on ostracods and polychaetes in the STR of the western North Pacific (Sassa & Kawaguchi, Reference Sassa and Kawaguchi2004). In the Gulf of Mexico, Centrobranchus nigroocellatus, Gonichthys cocco, M. obtusirostre, M. selenops and M. affine prey primarily on ostracods, and Lobianchia gemellarii and Hygophum taaningi preferred thaliaceans (Conley & Hopkins, Reference Conley and Hopkins2004). In this study, the gut contents during 1997 and 1998 from a broad area of the STR were examined, and no large differences were observed in the diet composition both between years and among areas, but a more detailed study is needed to clarify this.
Myctophid larvae are considered to actively select certain prey and that this selection changes with development (Sabatés & Saiz, Reference Sabatés and Saiz2000; Conley & Hopkins, Reference Conley and Hopkins2004; Sassa & Kawaguchi, Reference Sassa and Kawaguchi2005). Symbolophorus californiensis larvae fed mainly on copepod eggs and nauplii during the preflexion stage of ≤7.9 mm BL. The three subarctic species: D. theta, P. thompsoni, and T. taylori also fed on copepod nauplii during the larval stage in the transition region of the western North Pacific (Sassa & Kawaguchi, Reference Sassa and Kawaguchi2005). Since copepod eggs, due to the chorionic membrane, show a level of temporary resistance to digestion inside the digestive tract, their abundance may be overestimated in the early stage larvae (Conway et al., Reference Conway, McFadzean and Tranter1994). With growth, the importance of calanoid copepodites such as Pseudo/Paracalanus spp. increased greatly. In the largest size-class larvae of 16–22.2 mm BL, furcilia stage of euphausiids was also an important prey item. Although the larvae of D. theta, P. thompsoni and T. taylori fed largely on both Oithona spp. and Pseudo/Paracalanus spp. (Sassa & Kawaguchi, Reference Sassa and Kawaguchi2005), S. californiensis rarely consumed Oithona spp. Instead, S. californiensis larvae extensively consumed Pseudo/Paracalanus spp. that is also the main prey for larvae and juveniles of commercially caught fish of the Japanese anchovy and Pacific saury in the study area (Odate, Reference Odate1994; Sassa & Kawaguchi, Reference Sassa and Kawaguchi2005). Both Paracalanus spp. and Pseudocalanus spp. are numerically the most dominant small-sized copepods and are widely distributed in the epipelagic layer of the transition region, with the peak abundance from spring to early summer (Hattori, Reference Hattori1991; Odate, Reference Odate1994; Yamaguchi & Shiga, Reference Yamaguchi and Shiga1997). The main spawning season of S. californiensis is considered to be during April–May (Moser & Ahlstrom, Reference Moser, Ahlstrom and Moser1996; Sassa et al., Reference Sassa, Kawaguchi and Taki2007a). The peak abundance of S. californiensis larvae would correspond to the spring–summer peak in prey abundance in the STR (Sassa et al., Reference Sassa, Kawaguchi, Oozeki, Kubota and Sugisaki2004, Reference Sassa, Kawaguchi and Taki2007a).
Larvae of S. californiensis attain the largest size of any species of myctophid larvae in the western North Pacific (Moser & Ahlstrom, Reference Moser and Ahlstrom1970). Since routine metabolic rates of larval fish increase with body weight and habitat temperature (Giguère et al., Reference Giguère, Côté and St-Pierre1988; Finn et al., Reference Finn, Rønnestad, van der Meeren and Fyhn2002), and S. californiensis spend in the warm epipelagic layer throughout the larval stage (Tsukamoto et al., Reference Tsukamoto, Zenitani, Kimura, Watanabe and Oozeki2001; Sassa et al., Reference Sassa, Kawaguchi, Oozeki, Kubota and Sugisaki2004), the food requirements of this species would increase with larval growth. Although there was no change in the number of prey items consumed with growth in the larvae ≤11.9 mm BL, they showed a faster ontogenetic shift to larger prey, corresponding with a shift of major prey type with larval growth. Conversely, although there was no large increase in average prey size with growth in larvae ≥12 mm BL, the number of prey eaten was positively correlated with larval growth to meet their demand for energy. These ontogenetic changes would relate to enhanced foraging and capture abilities with growth, since swimming speed, capture success rates, and visual perception are functions of body length or age (Hunter, Reference Hunter and Lasker1981). Furthermore, our results suggest that the feeding strategy of S. californiensis larvae changed at approximately 12–14 mm BL when most fin rays have formed and are ossifying, and the stalked nature of the eyes becomes less apparent, i.e. most juvenile characteristics have developed, except for the photophores and pigments (Moser & Ahlstrom, Reference Moser and Ahlstrom1970).
The niche breadth of S. californiensis larvae also increased with growth. Although increases of size and number of prey items with larval growth have been reported for many other myctophids, the trophic niche breadth increase is a rather unusual phenomenon in this family (Sabatés & Saiz, Reference Sabatés and Saiz2000; Sassa & Kawaguchi, Reference Sassa and Kawaguchi2004, Reference Sassa and Kawaguchi2005). A similar ontogenetic shift in trophic niche breadth was reported only for Protomyctophum thompsoni, which also attains a large larval size of approximately 17–18 mm SL (Moser & Ahlstrom, Reference Moser and Ahlstrom1970; Sassa & Kawaguchi, Reference Sassa and Kawaguchi2005). Since the increase of niche breadth with development ensures the wide range of available food resources, it would be an effective mechanism for these two large-sized larvae to meet their metabolic requirements.
ACKNOWLEDGEMENTS
I thank Professor Emeritus K. Kawaguchi of the Ocean Research Institute, University of Tokyo, for valuable discussions during the course of this study. I am grateful to the captains, officers, and crews of the RV ‘Hakuho-Maru’, ‘Wakataka-Maru’, and ‘Soyo-Maru’ for their assistance in the field. Drs Y. Oozeki, H. Kubota, and H. Sugisaki of the National Research Institute of Fisheries Science kindly provided S. californiensis larvae for gut content analysis.
