Limnol.Oceanogr.,47(3),2002,730–741 (cid:113)2002,bytheAmericanSocietyofLimnologyandOceanography,Inc. Carbon sources for demersal fish in the western Seto Inland Sea, Japan, examined by (cid:100) (cid:100) 13C and 15N analyses Noriyuki Takai and Yasufumi Mishima National Institute of Advanced Industrial Science and Technology (AIST), Kure 737-0197, Japan Akemi Yorozu Graduate School of Biosphere Science, Hiroshima University, Higashi-Hiroshima 739-8528, Japan Akira Hoshika National Institute of Advanced Industrial Science and Technology (AIST), Kure 737-0197, Japan Abstract The relative importance of benthic and pelagic primary production for demersal fish and some invertebrates in thewesternSetoInlandSeaofJapanwasexaminedusingcarbonandnitrogenstableisotopeanalyses.Afewfishes, such as juvenile black rockfish Sebastes inermis and large Japanese anchovies Engraulis japonicus, had isotopic carbonsignaturessimilartopelagicparticulateorganicmatter((cid:50)20.1(cid:54)1.7‰),whichindicatesthattheirfoodwas derived fromproductioninthewatercolumn.However,92%ofthe401demersalfishthatwereanalyzedhad(cid:100)13C signatures((cid:50)17.0to(cid:50)13.0‰)similartothoseofbenthiccrustaceans,epilithicmicrophytobenthos,andmacroalgae andunlikethesignatureofpelagicparticulateorganicmatterorzooplankton.TheseresultssuggestthatintheSeto Inland Sea there is not a tight coupling between pelagic primary production and the food web of demersal fishes, but rather that these fishes are dependent on carbon from benthic primary production. Petersen and Curtis(1980)suggested thatpelagic–benthic areaoftheSetoInlandSeaaresimilartovaluesintheChuk- couplinginthemarineecosystemshouldbetighterathigher chiandnorthernBeringSeas(Seikietal.1985).Inaddition, latitudes than at lower latitudes. It appears that this latitu- macrobenthic biomass is only 56 g wet weight m(cid:50)2 in the dinal difference can be seen in the coastal shallow waters SetoInlandSea(Hashimotoetal.1997),incontrasttolevels along the North Pacific Ocean between the northernmost of 118–2,377 g wet weight m(cid:50)2 in the benthos oftheBering Arcticregionandthewesterntemperateregion.Inthehighly Shelf–Anadyr Water (Grebmeier et al. 1988). productiveChukchiandnorthernBeringSeaswhereprimary Here we address the relative importance of benthic pri- production ranges from 250 to 300 gC m(cid:50)2 yr(cid:50)1 (Walsh et mary production for the shallow coastal waters of Japan, al. 1989), a large portion of the organic matters being pro- where pelagic–benthic coupling is weak. In the Seto Inland duced in the water column falls ungrazed to the bottom Sea and arctic seas, benthic animals have been assumed to (Grebmeierand Barry 1991).Thedownwardorganiccarbon beprimarilydependentonwatercolumnprimaryproduction flux was reported to be 253–654 mgC m(cid:50)2 d(cid:50)1 in the north- (Tatara 1981), but recent studies showing the importance of ern Bering Sea (Fukuchi et al. 1993), and low surface sed- benthic primary production challenge this assumption. Ca- iment C:N ratios (wt.:wt.) of 6–8 were found under the hoon and Cooke (1992) demonstrated that the benthic pri- BeringShelf–AnadyrWater,whichsuggeststhatahighqual- maryproductioniscomparabletothewatercolumnprimary ity, nitrogen-rich organic material was being supplied to the production not only in macroalgal or seagrass beds but also benthic animals (Grebmeier et al. 1988). On the other hand, in so-called unvegetated habitats devoid of macrophytes. In in the shallow, productive Seto Inland Sea of Japan, surface the South Atlantic Bight, benthic microalgae contribute to sediment C:N ratios are relatively high (8–11) (Shinohara approximately40%ofthetotalprimaryproductionatdepths 1997; Mishima et al. 1999). This suggests that the quality of 14–40 m (Jahnke et al. 2000). We hypothesizedthatben- of the organic matter supplied to the benthos may be lower thic primary production might supply substantial organic than that in the Bering Sea, although annual primary pro- carbon to the demersal fish in the Seto Inland Sea. duction (280–460 gC m(cid:50)2) and the downward organic car- To test this hypothesis we analyzed the (cid:100)13C of the de- bon flux (140–440 mgC m(cid:50)2 d(cid:50)1) reported for the western mersalfishinthewesternSetoInlandSeatodeterminetheir carbon source. Since the (cid:100)13C of the animals increases only Acknowledgments about 1‰ through a prey–predator trophic link (DeNiroand WethankE.Wadaforhiscriticalreadingofthismanuscript.We Epstein 1978), the (cid:100)13C of the animals reflects the (cid:100)13C of aregratefultoM.UenoforhisusefulcommentsandadviceandH. primary producers atthe base ofthefoodweb.Thereported Yamaguchi for cooperation in the species identification of small average(cid:100)13Cvalueformarinephytoplanktonis(cid:50)22‰,low- crustaceans.WearealsogratefultoJ.J.Middelburgandtwoanon- erthanthe(cid:50)17‰signatureformarinebenthicalgae(France ymous reviewers for their valuable comments on the manuscript 1995), thus allowing the relative contribution of water col- and to W. Wurtsbaugh for English revision of themanuscript.This studywassupportedbyagrantfromJapanScienceandTechnology umn and benthic primary production to be estimated. In the Corporation. Chukchi Sea, the (cid:100)13C values of the demersal fish were re- 730 Carbon sources for demersal fish 731 24 m (Fig. 1). The bay is dotted with many islands in its central area, and thus its topography is very complicated.In particular, the Nasami Strait divides the bay into two semi- enclosed areas, northern and central Hiroshima Bay. The northern bay is one of the most eutrophic areas in the Seto Inland Sea and the polluted Ota River flows into this area from the northern shore. The chlorophyll a (Chl a) concen- tration increases and the dissolved oxygen becomes super- saturated in the surfacelayerofthenorthernbayinsummer, in contrast to the small seasonal change in the central bay (Hashimoto et al. 1994). Materials and methods Sampling—Fish were captured with a throw net on the island shore (Sta. Z) from 15 June 1999 to 30 June 2000, with a boat seine at Sta. Y ((cid:44)10 m in depth) on 21 October 1997 and 16 April 1998, and with a bottom trawl at Stas. H1–H4 (10–30 m in depth) on 14 May 1999 (Fig. 1). Stan- dardlengthsofthefisharereportedhere.Smallbenthiccrus- Fig. 1. Sampling locations in Hiroshima Bay. The isobathsare taceans including amphipods, an isopod, and decapods were shown with dotted lines for 20 m and broken lines for 100 m. collectedfromthesurfacesofstones,macroalgae,andshells of bivalves at Sta. Z. Mysids (Acanthomysis tenuicaudaand Acanthomysis spp.) and copepods that were extracted from portedtobeverysimilartothevaluesofzooplankton,clear- the stomach contents of Japanese horse mackerel Trachurus lyreflectingthetightpelagic–benthiccouplingthere(Dunton japonicus captured at Sta. H4 were also analyzed for stable et al. 1989; Schell et al. 1998). If water column primary isotope ratios. As supplementary data, bivalves, cephalo- production contributed significantly to the demersal fish in pods, large decapods, a starfish, holothurians, and an ascid- the Seto Inland Sea, the area’s demersal fish would show ianwerealsoanalyzedforstableisotoperatios.Wecollected depleted (cid:100)13C values like the fish in the Chukchi Sea. the Japanese oysterCrassostrea gigasfromseveraldifferent In this study, we also analyzed the nitrogen stableisotope kinds of habitat (the island shore [Sta. Z], the sea bottom ratio of the demersal fish as a secondary tracer. (cid:100)15N levels [Sta. H3], and the oyster raft [Sta. X]) in order to examine in both invertebrates and vertebrates show an increase of thehabitat-relateddifferencesinisotopicvalues.Theoysters about 3–4‰ per trophic level (Minagawa and Wada 1984), collected at Sta. X were hanging from a raft in the surface andthustheisotopicnitrogensignaturescanbeusedtomea- layer (0–6 m). sure trophic position. Consequently, schematic food web Particulate organic matter (POM) (0.7–125 (cid:109)m) for iso- structures depicted by (cid:100)13C–(cid:100)15N maps provide useful infor- topic analysis was collected from sea surface water at Stas. mation about the transport pathways of organic matter from H1–H4 on 14 May 1999 and filtered onto precombusted primary producers to top predators. (450(cid:56)C, 2 h) glass-fiber filters (Whatman GF/F type) after being sieved through a 125-(cid:109)m mesh sieve. The POM in Study area the surface water of area W was collected and processed from 10 August 1999 to 10 November 2000. We also sam- The Seto Inland Sea is the broad shallow area located to pled larger particles in the layer (cid:44)20 m on 14 May 1999 the southwest of the mainland of Japan (Fig. 1). It is 450 with a vertical haul of a 350-(cid:109)m mesh plankton net. These km long, with an area of 220,000 km2 and a mean depth of were sieved through 350-(cid:109)m and 125-(cid:109)m mesh sieves, and 37 m. The sea consists of several shallower semienclosed the stable isotope ratios of POM in each fraction were an- areas and deeper narrow channels, leading to the open sea alyzed. at three openings, the Bungo Channel, the Kii Channel, and Theepilithicorganicmatter(EOM)asanindicatorofepi- the Kanmon Strait. The latter strait leads to the Japan Sea lithic algae was collected at Sta. Z from 10 August 1999 to at its westernmost end, and is too narrow to affect the total 12 October 2000. Several submerged greenish stones were water exchange of the sea. The former two straits lead to collected at the sublittoral fringe at low tides of spring tide, the Pacific Ocean at the southwestern (the Bungo Channel) and the surfaces of the stones were brushed in the sea water and the southeastern (the Kii Channel) end of the sea. The that had been filtered through GF/F filters. The brushedma- mass of water exchanged through the Bungo Channel is terial from the biofilm was passed through a 125-(cid:109)m sieve twice that through the Kii Channel (Fujiwara 1983), andthe and then filtered onto GF/F filters for the isotopic analyses residence time of the total water in the sea is assumed to be of EOM of 0.7–125 (cid:109)m. This EOM collected at the sublit- less than several years (Takeoka 1984). toral fringe consisted mainly of diatoms and detritus (Takai HiroshimaBayislocatedtothenorthoftheBungoChan- unpubl. data). Using the same procedure, we also analyzed nel and has an area of ca. 1,000 km2 and a mean depth of thegreenorganicmatter(0.7–125(cid:109)m)attachedtotheshells 732 Takai et al. Table1. Thecarbon(upper)andnitrogen(lower)stableisotoperatios(‰)ofPOM,mysids(AcanthomysistenuicaudaandAcanthomysis spp.), copepods, and SOM collected on 14 May 1999. The seasonal means of POM in the surface layer of area W from 14 May 1999 to 10 Nov 2000 and EOM at the sublittoral fringe of Sta. Z from 10 Aug 1999 to 12 Oct 2000 are also shown. POM* EOM Location 0.7–125 (cid:109)m 125–350 (cid:109)m (cid:46)350 (cid:109)m Mysids† Copepods† SOM 0.7–125 (cid:109)m 14 May 1999 H1‡ (cid:50)22.3 (cid:50)22.0 — (cid:50)22.2 9.5 10.7 — 7.2 H2‡ (cid:50)21.3 — — (cid:50)20.7 9.2 — — 7.7 H3 (cid:50)21.9 (cid:50)22.6 (cid:50)22.0 (cid:50)20.6 8.5 10.6 10.4 7.4 H4 (cid:50)22.2 (cid:50)22.6 (cid:50)22.1 (cid:50)21.0 (cid:50)20.8 (cid:50)20.2§ 8.2 9.3 10.2 8.4 8.5 6.9 The seasonal mean ((cid:54)SD) W (cid:50)20.1(cid:54)1.7 8.3(cid:54)1.3 (n (cid:53) 10) Z (cid:50)15.4(cid:54)1.8 8.4(cid:54)1.4 (n (cid:53) 7) *ThePOMsof 0.7–125(cid:109)mwerecollectedfromtheseasurface.ThePOMsof125–350(cid:109)mand(cid:46)350(cid:109)mwerecollectedfromthedepthlayerof0–20 mwithaplanktonnet. †ThesewereextractedfromthestomachofJapanesehorsemackerelTrachurusjaponicusandweredefatted. ‡Dashindicatesnodata. §TheSOMatSta.H4wascollectedon27Oct1998. of Japanese oysters that were collected from the sea bottom tios,(cid:100)13C and (cid:100)15N,are expressedaspermildeviationsfrom (Sta. H3) and the oyster raft (Sta. X). the standard as defined by the following equation: The surface sedimentary organic matter (SOM) was col- (cid:100)13C, (cid:100)15N (cid:53) [R /R (cid:50) 1] (cid:51) 1,000 (‰) lected with a bottom sampler at Stas. H1–H3 on 14 May sample standard 1999. At Sta. H4, additional sediment was collected on 27 where R (cid:53) 13C/12C or 15N/14N. Belemnite (PDB) and atmo- October1998andanalyzedforisotoperatios,sincethesam- sphericnitrogenwereusedastheisotopestandardsofcarbon ple collection at this station at the trawling date was unsuc- and nitrogen, respectively. The analytical precision for the cessful. isotopic analyses was (cid:35)0.28‰ for both (cid:100)13C and (cid:100)15N. Stable isotope analyses—The samples were stored at Results (cid:50)20(cid:56)C. We analyzed the muscle tissues of fish, bivalves, cephalopods,largedecapods,a starfish,holothurians,andan Stable isotopic distribution of EOM, POM, and SOM— ascidian. The muscles were excised from the trunk behind a The (cid:100)13C of EOM (0.7–125 (cid:109)m) at the sublittoral fringe of pectoralfininfish,fromtheadductormuscleinthebivalves, the island shore (Sta. Z) averaged (cid:50)15.4 (cid:54) 1.8‰ ((cid:54)SD) from the mantle in the cephalopods,and fromthetube-foots from August 1999 to October 2000 (n (cid:53) 7) (Table 1, Fig. in the starfish. The isotopic ratios of small crustaceans were 2). Likewise the (cid:100)13C of the organic matter (0.7–125 (cid:109)m) analyzed for mixed individuals except Melita rylovae, attached to the shells of the Japanese oysters from the sea Cleantiellaisopus,andHeptacarpusfutilirostris.Theanimal bottom (10–30 m in depth) had enriched (cid:100)13C values of tissues were dried at 60(cid:56)C and ground to a fine powder and (cid:50)13.9‰ and those from the oyster raft in the surface layer lipids were removed with a chloroform:methanol (2:1) so- had enrichments of (cid:50)15.6‰ (Fig. 2). These values were lution. The POM and EOM samples were exposed to the clearly more enriched relative to the average (cid:100)13C of POM vapor of concentrated HCl for a day in order to eliminate (0.7–125 (cid:109)m) collected in the surface layer of the central carbonates and then were dried in a vacuum desiccator. The bay (area W) from May 1999 to November 2000 (n (cid:53) 10, SOM samples were saturated with 1 M HCl solution for a (cid:50)20.1 (cid:54) 1.7‰; Table 1, Fig. 2). This (cid:100)13C difference be- day in order to eliminate carbonates and then were dried on tweenattachedorganicmatterandPOMwasconsistentwith a hot plate. We analyzed one sample from each sampling the general characteristics of the benthic–planktonic differ- date for POM, EOM, SOM, and the shell-attached organic ence in isotopic ratios compiled by France (1995). matter. The (cid:100)15N of EOM (0.7–125 (cid:109)m) at Sta. Z averaged 8.4 Stable isotope ratios of carbon and nitrogen were mea- (cid:54) 1.4‰ (n (cid:53) 7), which is very similar to the average (cid:100)15N sured with a MAT 252 mass spectrometer (Finnigan MAT) of 8.3 (cid:54) 1.3‰ (n (cid:53) 10) in offshore POM (0.7–125 (cid:109)m) in coupled with an element analyzer (Carlo Erba). Isotope ra- area W (Table 1; Fig. 2). The organic matter attached to Carbon sources for demersal fish 733 the demersal fishes, whereas the depleted signatures of off- shore POM did not (Table 1; Fig. 2). The (cid:100)15N of the fish averaged 14.6 (cid:54) 0.5‰, rangingfrom 13.1‰ in black rockfish to 15.8‰ in spottybelly greenling Hexagrammosagrammus(Table2).Thesevaluesweremuch more enriched than those of the EOM (6.5–9.9‰) and of the small benthic crustacean (6.7–11.8‰; Tables 1, 3). The young-of-the-year black rockfish sampled in June showed clearly depleted isotopic values for both (cid:100)13C and (cid:100)15N ((cid:50)18.6 to (cid:50)17.0‰ in (cid:100)13C and 13.1 to 14.4‰ in (cid:100)15N; Fig. 3). The (cid:100)15N of the rockfish increased linearly with length,consistentwithsize-related(cid:100)15Nchangesreportedfor carnivorous fish species (Takai and Sakamoto 1999), while the (cid:100)13C increased in a stepwise manner for fish between62 and 66 mm. There were significant correlations betweenthe fish length and the stable isotope ratios (n (cid:53) 52; (cid:100)13C, r (cid:53) 0.74, P (cid:44) 0.0001; (cid:100)15N, r (cid:53) 0.85, P (cid:44) 0.0001). Fig. 2. The carbon (square) and nitrogen (circle)stableisotope Stable isotope ratios of fish in the deeper area—The(cid:100)13C ratios (‰) of POM (0.7–125 (cid:109)m) in the surface layer in area W of fish in the deeper area (Stas. Y, H1, H2, H3, and H4) and of EOM (0.7–125 (cid:109)m) at the sublittoral fringe of Sta. Z. The rangedfrom(cid:50)19.5‰inJapaneseanchovyEngraulisjapon- valuesofshell-attachedorganicmatterarealsoshownforJapanese icus(Sta.Y)to(cid:50)12.6‰insurffishDitrematemmincki(Sta. oysters Crassostrea gigas from the sea bottom of Sta. H3 and the Y) (Table 4). The fish (cid:100)13C was distributed from (cid:50)17.0 to oyster raft of Sta. X on 10 November 2000. (cid:50)13.0‰ in 203 (96%) of the 212 samples and was very similar to the (cid:100)13C distribution of the fish at Sta. Z (Table shells(0.7–125(cid:109)m)wasslightlyenrichedin(cid:100)15Nwith11.3– 2). Likewise, benthicinvertebratesinthedeeperareamostly 12.0‰ relative to the EOM and POM (Fig. 2). had enriched (cid:100)13C values of (cid:50)17.0 to (cid:50)13.0‰ (Table 5). In On 14 May 1999, the local differences and the particle- particular, the Japanese oysters showed enriched(cid:100)13Cvalues size–related differences in the POM isotopic values were of (cid:50)15.7 to (cid:50)14.9‰, in contrast to the (cid:100)13C distribution of examined (Table 1). Irrespective of the sampling stationand other filter-feeding benthos, the egg cockles Fulvia mutica, particle size, the POMs had depleted (cid:100)13C ranging from and the ascidian Styela plicata. (cid:50)22.6 to (cid:50)21.3‰. These (cid:100)13C values were very similar to The (cid:100)15N of fish in the deeper area ranged from 7.4‰ in the (cid:100)13C of (cid:50)21.0 to (cid:50)20.8‰ in the mysids (Acanthomysis Japanese anchovy (Sta. Y) to 17.5‰ in cutlassfish Trichiu- tenuicauda and Acanthomysis spp.) and the copepods ex- rus japonicus (Sta. H4) (Table 4), showing much more var- tracted from the stomach of the Japanese horse mackerel at iability relativetothefishinthelittoralareas(Sta.Z)(Table Sta.H4.Ontheotherhand,the(cid:100)15NofPOMclearlyshowed 2).The(cid:100)15Ndistributionofbenthicinvertebratesinthedeep- both local and particle-size–related differences (Table 1). er area also had highly variable (cid:100)15N distributions, with val- The(cid:100)15NofPOM(cid:46)125(cid:109)mwasmoreenriched(9.3–10.7‰) ues ranging from 9.5‰ in Japanese trepang Apostichopus relative to the POMs of 0.7–125 (cid:109)m (8.2–9.5‰), and the japonicus (Sta. H2) to 17.9‰ in Japanese squid Loliolus (cid:100)15N of POM increased northward from 8.2‰ (Sta. H4) to japonica (Sta. H2) (Table 5). 9.5‰ (Sta. H1) in 0.7–125-(cid:109)m particles and from 9.3‰ In two fish species, cardinal fish Apogon lineatus and (Sta. H4) to 10.7‰ (Sta. H1) in 125–350-(cid:109)m particles. The young-of-the-year gurnard Lepidotrigla microptera (48–76 (cid:100)15N values of 8.4–8.5‰ in the mysids and copepods were mm), clear local variations in (cid:100)15N were found (Fig. 4). The similar to the (cid:100)15N of 0.7–125 (cid:109)m POM. (cid:100)15N of these fishes was significantly more enriched in the The (cid:100)13C of the SOM ranged from (cid:50)22.2to(cid:50)20.2‰and northernbaythaninthecentralbay.The(cid:100)15Nofthecardinal was similar to the values of the POM (Table 1). The SOM fish was 16.7 (cid:54) 0.4‰ (n (cid:53) 4) at Sta. H1 and 15.0 (cid:54) 0.3‰ from the mouth of the Ohta River was slightly depleted in (n (cid:53) 41) at Stas. H3 and H4 (Mann–Whitney’s U-test; P (cid:53) (cid:100)13C((cid:50)22.2‰,Sta.H1).The(cid:100)15NvariationintheSOMwas 0.001). The (cid:100)15N of the young gurnard was 15.7 (cid:54) 0.3‰ (n small, ranging from 6.9 to 7.7‰. (cid:53) 17) at Sta. H1 and 14.6 (cid:54) 0.4‰ (n (cid:53) 9) at Sta. H4 (Mann–Whitney’s U-test; P (cid:44) 0.0001). Stableisotoperatios offish in theislandshore—The(cid:100)13C Four large Japanese anchovies (cid:36)125 mm showed pecu- of fish in the island shore (Sta. Z) ranged from (cid:50)18.6‰ in liarly depleted isotopic values in both (cid:100)13C and (cid:100)15N (Fig. black rockfish Sebastes inermis to (cid:50)13.8‰ in motleystripe 5).Theseanchoviesrangedfrom(cid:50)19.5to(cid:50)18.1‰for(cid:100)13C rainbowfishHalichoerestenuispinnis(Table2).Thefish(cid:100)13C and from 7.4 to 9.4‰ for (cid:100)15N and were in contrast to the was distributed from (cid:50)17.0 to (cid:50)13.0‰ in 166 (88%)ofthe enrichmentsofanchovies(cid:35)120mmwith(cid:50)17.2to(cid:50)13.1‰ 189 samples and was similar to the (cid:100)13C values of (cid:50)17.3 to for (cid:100)13C and 12.7 to 16.6‰ for (cid:100)15N. (cid:50)12.5‰ in the small benthic crustaceansthatarethoughtto betheprincipalpreyforthesefish(Table3).The(cid:100)13Cvalues The (cid:100)13C–(cid:100)15N diagram for the Hiroshima Bay food derivedfrombenthicprimaryproductionfromEOMandthe web—Thefishanalyzedinthisstudyweredividedintothree shell-attached organic matter overlapped those of most of distinctgroupsonthe(cid:100)13C–(cid:100)15Ndiagram(Fig.6).Onegroup 734 Takai et al. Table 2. The carbon (upper) and nitrogen (lower) stableisotoperatios(‰; mean (cid:54) SD) offishcapturedwithathrownetintheisland shore (Sta. Z) in Hiroshima Bay. The sample number (left) and standard length (mm; right) are shown in parentheses. Species 15 Jun 99 10 Aug 99 26 Oct 99 22 Feb 00 7 Mar 00 30 Jun 00 Perciformes Surf fish Ditrema temmincki (cid:50)14.9 (cid:50)14.1 (cid:50)17.3(cid:54)0.1 15.0 15.2 13.8(cid:54)0.2 (1; 52) (1; 123) (5; 62–69) Richardson’s dragonet Repomucenusrichardsonii (cid:50)15.5(cid:54)0.3 13.9(cid:54)0.4 (3; 62–80) Motleystripe rainbowfish Halichoerestenuispinnis (cid:50)14.7(cid:54)0.5 (cid:50)14.8(cid:54)0.4 (cid:50)15.5(cid:54)0.6 (cid:50)15.3(cid:54)0.3 14.8(cid:54)0.3 14.3(cid:54)0.2 14.2(cid:54)0.3 14.3(cid:54)0.2 (20; 50–90) (10; 59–102) (8; 54–94) (10; 75–84) Pudding wife Halichoerespoecilopterus (cid:50)14.9(cid:54)0.4 (cid:50)15.5(cid:54)0.3 (cid:50)14.1 (cid:50)15.6 15.0(cid:54)0.3 14.5(cid:54)0.3 14.2 14.6 (13; 54–112) (4; 58–99) (1; 83) (1; 120) Scorpaeniformes Spottybelly greenling Hexagrammosagrammus (cid:50)15.0(cid:54)1.1 (cid:50)15.5(cid:54)0.2 (cid:50)15.9(cid:54)0.3 (cid:50)14.5(cid:54)0.1 (cid:50)14.5(cid:54)0.1 15.1(cid:54)0.7 14.5(cid:54)0.2 14.6(cid:54)0.2 14.7(cid:54)0.6 14.9(cid:54)0.9 (3; 66–146) (6; 81–102) (3; 90–141) (6; 43–135) (3; 52–136) Tiny stinger Hypodytes rubripinnis (cid:50)15.2(cid:54)0.7 (cid:50)14.8(cid:54)0.2 (cid:50)14.0 (cid:50)14.4 (cid:50)14.8(cid:54)0.2 14.4(cid:54)0.5 14.7(cid:54)0.4 15.3 14.6 14.8(cid:54)0.2 (8; 35–70) (10; 50–59) (1; 63) (1; 52) (7; 44–66) Black rockfish Sebastes inermis (cid:50)17.7(cid:54)0.7 (cid:50)17.1, (cid:50)16.8 (cid:50)18.0(cid:54)0.3 (62 mm (cid:36) SL) 14.1(cid:54)0.2 13.8, 14.1 13.7(cid:54)0.4 (6; 45–53) (2; 51, 55) (11; 41–62) Black rockfish Sebastes inermis (cid:50)15.6(cid:54)0.3 (cid:50)15.2(cid:54)0.3 (cid:50)14.9(cid:54)0.2 (cid:50)15.9, (cid:50)15.2 (66 mm (cid:35) SL) 15.2(cid:54)0.2 14.8(cid:54)0.3 15.0(cid:54)0.2 14.9, 15.2 (10; 107–127) (14; 66–115) (5; 78–118) (2; 91, 122) Tetraodontiformes Net-work filefish Rudarius ercodes (cid:50)16.0(cid:54)0.9 (cid:50)15.9(cid:54)0.3 14.7(cid:54)0.5 14.6(cid:54)0.3 (4; 33–39) (10; 38–44) (A) consisted of only young-of-the-year black rockfish at pleted (cid:100)13C values ranging from (cid:50)18.6 to (cid:50)17.0‰ in both Sta. Z in June. These fish were distributed intherangefrom 1999 and 2000, in contrast to enriched values of (cid:50)16.0 to (cid:50)18.6 to (cid:50)17.0‰ in (cid:100)13C and from 13.1 to 14.4‰ in (cid:100)15N. (cid:50)14.6‰ in the larger rockfish (cid:36)66 mm (Table 2, Fig. 3). Group B consisted only of the large Japanese anchovies Stomach content analysis has shown that the rockfish feed (cid:36)125 mm. Both (cid:100)13C and (cid:100)15N in these fish were strikingly mainly on copepods at the planktonic larval stage in winter depleted, with (cid:50)19.5 to (cid:50)18.1‰ in (cid:100)13C and from 7.4 to andonbenthiccrustaceansafterthefishsettleinspring(Har- 9.4‰ in (cid:100)15N. Group C includes all the other fish collected ada1962).Thisindicatesthatthe13Cdepletionofplanktonic in both the island shore and deeper area. The isotopic ratios rockfish does reflect their diet of zooplankton and the (cid:100)13C ofthesefishweremostlydistributedfrom(cid:50)17.0to(cid:50)13.0‰ isotopic ratios of (cid:50)20.1 (cid:54) 1.7‰ of water column primary in (cid:100)13C and from 13.0 to 18.0‰ in (cid:100)15N. producers. We assumed that the rockfish would changetheir carbon source from water column primary production to Discussion benthic primary production with their life history change in feeding habit. However, the two young-of-the-year black Carbon source and trophic position of fish dependent on rockfish collected at Sta. H1 on 14 May 1999 showed en- water column primary production—All the young-of-the- richedisotopicvaluesof(cid:50)16.4to(cid:50)15.3‰in(cid:100)13Cand14.4 year black rockfish at Sta. Z in June showed peculiarly de- to 14.7‰ in (cid:100)15N (Table 4, Fig. 3). These young inhabiting Carbon sources for demersal fish 735 Table 3. The carbon (left) and nitrogen (right) stable isotope ratios (‰) of crustaceans in the island shore (Sta. Z) in Hiroshima Bay. A mixtureof some samples wereanalyzed for thecrustaceans.Themixedsamplenumber(left)andbodylength(mm;right)areshownin parentheses. Species 10 Aug 99 26 Oct 99 22 Feb 00 7 Mar 00 30 Jun 00 30 Aug 00 12 Oct 00 Amphipoda Gammaridea Ampithoe lacertosa (cid:50)15.7; 6.7 (cid:50)17.3; 7.4 (cid:50)13.5; 7.5 (8; 5.5–13.2) (10; 11.6–18.8) (10; 11.4–14.0) Pontogeneia rostrata (cid:50)16.7; 7.9 (20; 3.2–5.1) Melita koreana (cid:50)13.1; 9.2 (cid:50)14.2; 8.1 (cid:50)12.5; 8.1 (20; 5.8–8.2) (10; 4.5–7.0) (10; 5.0–7.5) Melita rylovae (cid:50)14.6; 7.7 (1; 9.0) Hyale sp. 1 (cid:50)14.2; 8.6 (cid:50)14.0; 8.7 (10; 7.8–11.4) (10; 10.8(cid:50)12.0) Hyale sp. 2 (cid:50)15.0; 8.5 (cid:50)14.7; 8.3 (7; 6.2–9.5) (10; 8.0–9.7) Podocerus sp. (cid:50)16.8; 8.8 (19; 2.0–6.5) Caprellidea Caprella penantis (cid:50)15.2; 9.8 (10; 5.9–9.0) Isopoda Cleantiella isopus (cid:50)13.5; 9.5 (1; 16.9) Decapoda Heptacarpus futilirostris (cid:50)15.8; 11.8 (1; 12.4) Eualus sinensis (cid:50)16.9; 11.5 (cid:50)15.7; 11.4 (2; 6.0, 10.0) (2; 8.4, 10.0) the deeper area might have settled the bottom and changed differences between the ascidian and POMs at Sta. H1 were their feeding habit much earlier relative to the young in the smaller: 3.9‰ for 0.7–125-(cid:109)m particles and 2.7‰ for 125– shore. 350-(cid:109)m particles relative to the differences in the young Depleted (cid:100)13C values of (cid:50)19.0 to (cid:50)17.0‰ were also rockfish. The ascidian may feed not only on zooplankton, found for two kinds of filter-feeding benthos, egg cockles but also on phytoplankton. (Sta. H4) and an ascidian (Sta. H1), ranging from (cid:50)18.2 to The egg cockles at Sta. H4 showed peculiarly depleted (cid:50)17.7‰ (Table 5). As expected, these benthic filter feeders (cid:100)15N values of 10.0 (cid:54) 0.1‰ (range: 9.8 to 10.1‰; n (cid:53) 4) appear to depend on water column primary production. in contrast to the young rockfish and the ascidian (Table 5). The (cid:100)15N of the young-of-the-year black rockfish in June These (cid:100)15N values were as depleted as the value (10.2‰) of averaged13.8(cid:54)0.4‰(range:13.1to14.4‰;n(cid:53)17;Table POM ((cid:46)350 (cid:109)m) at Sta. H4 (Table 1). Consequently, they 2,Fig.3).Thisaveragevaluewas3–4‰moreenrichedthan likely feed on 15N-depleted components of POMs, such as the POMs of (cid:46)350 and 125–350 (cid:109)m, and 5–6‰ more en- phytoplankton. riched than the POMs of 0.7–125 (cid:109)m, mysids, or copepods fromthecentralbay(Stas.H3andH4).Consideringthatthe Carbon sourceandtrophicpositionofmigratorsfromthe (cid:100)15N increase per trophic level in the shallow water near open sea—Japanese anchovies (cid:36)125 mmshowedapeculiar Japanwasreportedtobe3.4(cid:54)1.1‰(EastChinaSea;Min- isotopic distribution (Table 4, Fig. 5). The 20 anchovies agawa and Wada 1984), the 15N-depleted POMs (0.7–125 (cid:35)120 mm had mean (cid:100)13C signatures of (cid:50)15.5 (cid:54) 1.2‰ and (cid:109)m), mysids, and copepods are too depleted in (cid:100)15N to be mean (cid:100)15N signatures of 14.4 (cid:54) 1.0‰, similar to the adult the main food source for the young rockfish. It is likely that anchovyintheSagamiBay(PacificsideofthecentralJapan) some kinds of zooplankton composing the POMs (cid:46)125 (cid:109)m in October 1993 ((cid:50)17 to (cid:50)15‰ for (cid:100)13C and 11 to 14‰ would be the main food source for the young rockfish. for (cid:100)15N; Lindsay et al. 1998), while the four anchovies The (cid:100)15N of 13.4‰ in the ascidian collected at Sta. H1 (cid:36)125 mm had averages of (cid:50)18.8 (cid:54) 0.6‰ in (cid:100)13C and 8.5 was very similar to the values of the young rockfish, and (cid:54) 0.8‰ in (cid:100)15N. The latter were similar to the anchovies thus it appears as if their trophic positions were similar (Ta- captured in the western North Pacific Ocean offJapan((cid:50)20 ble5).However,theascidianlikelyfeedsonslightlysmaller to (cid:50)18‰ in (cid:100)13C and 7 to 10‰ in (cid:100)15N; Mitani2000).Con- componentsinPOMsthantheyoungrockfish,sincethe(cid:100)15N sidering that most populations of the anchovy are supposed 736 Takai et al. lected from temperate coastal waters (cid:44)10 m of the Japan Sea. Accordingly,the enriched(cid:100)13Cvaluesofthesmallben- thic crustaceans were inferred to be ubiquitous not only at the sublittoral fringe but also in the deeper area (up to at least 10 m in depth). These 13C-enriched fish and small benthic crustaceans overlappedthe(cid:100)13Csignaturesofepilithicandshell-attached organic matter but not the depleted signatures of planktonic POM (Table 1, Fig. 2). Table 6 shows reference data of the isotopicratiosofmacroalgae(nondefatted;Takaietal.2001) and seagrasses Zostera marina (nondefatted; Takai unpubl. data). The (cid:100)13C signatures of the seagrasses were clearly moreenriched thanthevaluesoftheanimals,whilethe(cid:100)13C distribution of the macroalgae overlapped with that of the 13C-enriched animals. Based on the overlapping (cid:100)13C distributions of the ma- croalgae, epilithic organic matter, and the 13C-enriched ani- mals, it appears that the organisms receive a substantialcar- bon supply from the benthic primary producers. Quantitatively important carbon transport from microphyto- benthos to macrobenthic animals was recently confirmed in a tidal flat ecosystem with a stable isotope tracer-addition experiment and by analyzing natural isotope abundances (Herman et al. 2000; Middelburg et al. 2000). The resultsin our study indicate that such benthic carbon transport would be important not only in the shore ecosystem but also in the 10–30 m depth strata. Similarly enriched (cid:100)13C values were also found for most benthic invertebrates in the deeper area of Hiroshima Bay, which suggests substantial carbon supply from the benthic primary production (Table 5, Fig. 6). In particular, all the Fig.3. Therelationshipbetweenstandardlengthandstableiso- Japanese oysters showed enriched (cid:100)13C values of (cid:50)15.7 to tope ratios in black rockfish Sebastes inermis collected at Stas. Z (cid:50)14.9‰, in contrast to the (cid:100)13C distribution of the other and H1 (n (cid:53) 52). (A) (cid:100)13C. (B) (cid:100)15N. filter-feeding benthos, the egg cockles,andtheascidian.Ac- cording to Riera and Richard (1996), oysters in an estuarine to migrate into the Seto Inland Sea from the open sea in bay in France were also enriched in (cid:100)13C, which suggests springandleavethereinwinter(Takao1985),theextremely that they were feeding on microphytobenthos from an ad- depleted (cid:100)13C and (cid:100)15N values of the anchovies in April are jacent wide mudflat with signatures of about (cid:50)16‰. It is interpretedasfeedingrecordsofthepelagicdietsintheopen likely that the enriched (cid:100)13C values of the oysters in Hiro- sea during the winter. This means that the isotopic valuesof shima Bay and the estuarine oysters in France indicate that these 13C-depleted anchovies would reflect an oceanic water theyarefeedingonmicrophytobenthos.Hereweneedtopay column carbon supply. attention to the enriched (cid:100)13C values of (cid:50)15.2 (cid:54) 0.3‰ in Japanese anchovies have generally been assumed to feed the oysters from the raft at Sta. X (Table 5). Sincethewater primarily on copepods, larvae, and diatoms (Kondo 1971), depthatSta.Xwasabout30m,itisunlikelythattheoysters but the stomach content analysis of anchovies captured in hanging in the upper 6 m of the water column would feed tidelandsofthewesternSetoInlandSeaatfloodtideshowed onthemicroalgaeinhabitingtheseabottom.Wesuggestthat that they feed not only on plankton but also on large num- the enriched (cid:100)13C values of those oysters might be derived bers of benthic animals including decapods, mysids, amphi- from the shell-attached organic matter, which was similarly pods,andpolychaetes(Zinnouchi1977).Itisthuslikelythat enriched ((cid:50)15.6‰) (Table 5). diversefeedingoftheanchoviesduringtheirstayintheSeto The average (cid:100)15N (11.6 (cid:54) 0.2‰) of the small decapods Inland Sea increased both their (cid:100)13C and (cid:100)15N signatures. at Sta. Z was 3.2‰ more enriched than the average of the EOM (8.4 (cid:54) 1.4‰), 3.0‰ more enriched than that of ma- Carbon source and trophic position of fish depending on croalgalChlorophyceae(8.6(cid:54)0.5‰;Takaietal.2001),and the benthic primary production—The (cid:100)13C signatures of 3.1‰ more enriched than that of Phaeophyceae (8.5 (cid:54) most fish and small crustaceans collected from both the 1.0‰; Takai et al. 2001) (Table 3, Fig. 6).Thesedifferences shore and the deeper area were distributed from (cid:50)17.0 to were consistent with the (cid:100)15N increase per trophic level in (cid:50)13.0‰ (Tables 2, 3, 4, Fig. 6). Although small benthic the East China Sea (3.4 (cid:54) 1.1‰; Minagawa and Wada crustaceans were not collected in the deeper area in Hiro- 1984), and accordingly the small decapods were estimated shimaBay,Yamaguchi(2000)foundsimilarlyenriched(cid:100)13C to be typical primary consumers dependent on the organic values of amphipods, isopods, and decapods in samplescol- matter originating from the benthic primary producers. Carbon sources for demersal fish 737 Table4. Thecarbon(upper)andnitrogen(lower)stableisotoperatios(‰;mean(cid:54)SD)offishcapturedwithaboatseine(Sta.Y)and a bottom trawl (Stas. H1–H4) in Hiroshima Bay. The sample number (left) and standard length (mm; right) are shown in parentheses. 21 Oct 1997 16 Apr 1998 14 May 1999 Species Y Y H1 H2 H3 H4 Clupeiformes Gizzard shad Konosirus punctatus (cid:50)15.1 (cid:50)14.5(cid:54)0.2 (cid:50)14.6(cid:54)0.1 (cid:50)14.7(cid:54)0.8 (cid:50)14.5(cid:54)0.3 14.7 15.3(cid:54)0.5 14.6(cid:54)0.7 14.3(cid:54)0.5 14.8(cid:54)0.7 (1; 230) (7; 152–228) (3; 220–230) (11; 207–232) (20; 203–238) Japanese anchovy Engraulis japonicus (cid:50)14.3(cid:54)0.9 (cid:50)16.4, (cid:50)15.3 (cid:50)16.3(cid:54)0.5 (cid:50)16.4(cid:54)0.7 (120 mm (cid:36) SL) 14.4(cid:54)0.7 12.7, 15.0 14.8(cid:54)1.1 14.0(cid:54)1.2 (8; 110–120) (2; 110, 115) (6; 69–91) (4; 74–114) Japanese anchovy Engraulis japonicus (cid:50)18.8(cid:54)0.6 (125 mm (cid:35) SL) 8.5(cid:54)0.8 (4; 125–140) Perciformes Cardinal fish Apogon lineatus (cid:50)14.8(cid:54)0.1 (cid:50)15.1 (cid:50)15.8(cid:54)0.3 16.7(cid:54)0.4 15.2 15.0(cid:54)0.3 (4; 43–53) (1; 46) (40; 32–56) Japanese horse mackerel Trachurus japonicus (cid:50)15.3, (cid:50)14.4 (cid:50)15.3(cid:54)0.1 16.0, 16.0 15.1(cid:54)0.0 (2; 160, 180) (3; 150–155) Ponyfish Leiognathus nuchalis (cid:50)15.4(cid:54)0.3 (cid:50)15.2(cid:54)0.4 15.8(cid:54)0.4 16.1(cid:54)0.6 (20; 78–117) (20; 78–115) Surf fish Ditrema temmincki (cid:50)12.6 (cid:50)14.7(cid:54)0.1 16.7 14.6(cid:54)0.1 (1; 200) (3; 152–180) Cutlassfish Trichiurus japonicus (cid:50)14.6 17.5 (1; 700)* Scorpaeniformes Black rockfish Sebastes inermis (cid:50)16.4, (cid:50)15.3 14.4, 14.7 (2; 38, 51) Gurnard Lepidotrigla microptera (cid:50)15.0(cid:54)0.1 (cid:50)16.6(cid:54)0.4 (76 mm (cid:36) SL) 15.7(cid:54)0.3 (cid:50)14.6(cid:54)0.4 (17; 56–76) (9; 48–68) Gurnard Lepidotrigla microptera (cid:50)14.7, (cid:50)14.3 (cid:50)14.2(cid:54)0.4 (cid:50)14.3, (cid:50)14.0 (cid:50)14.3 (141 mm (cid:35) SL) 16.6, 16.8 16.3(cid:54)1.1 15.8, 17.3 15.6 (2; 152, 164) (4; 141–198) (2; 142, 157) (1; 190) Tetraodontiformes Net-work filefish Rudarius ercodes (cid:50)15.0(cid:54)0.5 15.1(cid:54)0.6 (6; 31–48) Finepatterned puffer Takifugu poecilonotus (cid:50)15.0(cid:54)0.9 (cid:50)14.3 14.4(cid:54)0.3 15.8 (7; 59–85) (1; 84) *ThetotallengthwasmeasuredforTrichiurusjaponicus. 738 Takai et al. Table5. Thecarbonandnitrogenstableisotoperatios(‰;mean(cid:54)SD)ofbenthicinvertebratesinHiroshimaBay.Theisotopicvalues of shell-attachedorganic matter (0.7–125 (cid:109)m) from Japaneseoysters are also shown in parentheses. Species Sampling date Station Depth (m) n (cid:100)13C (cid:100)15N Mollusca Bivalvia Japanese oyster Crassostreagigas Shore 22 Feb 00 Z — 13 (cid:50)15.4(cid:54)0.2 11.0(cid:54)0.3 Sea bottom 10 Nov 00 H3 10–30 2 (cid:50)15.6, (cid:50)15.4 12.6, 12.9 [(cid:50)13.9] [11.3] Oyster raft 10 Nov 00 X 0–6 5 (cid:50)15.2(cid:54)0.3 12.9(cid:54)0.2 [(cid:50)15.6] [12.0] Egg cockle Fulvia mutica 14 May 99 H4 10–30 4 (cid:50)17.9(cid:54)0.2 10.0(cid:54)0.1 Cephalopoda Kobi cuttlefish Sepia kobiensis 14 May 99 H1 10–30 1 (cid:50)13.7 14.3 Golden cuttlefish Sepia esculenta 14 May 99 H4 10–30 1 (cid:50)13.6 14.8 Japanese squid Loliolus japonica 14 May 99 H2 10–30 2 (cid:50)14.3, (cid:50)14.1 16.8, 17.9 Long-armed octopus Octopus minor 14 May 99 H4 10–30 1 (cid:50)14.4 15.2 Arthropoda Crustacea(Decapoda) Mantis shrimp Oratosquilla oratoria 14 May 99 H1 10–30 1 (cid:50)14.7 15.8 Myra fugax (Leucosiidae) 14 May 99 H1 10–30 1 (cid:50)14.2 16.0 Echinodermata Asteroidea Starfish Asterias amurensis 14 May 99 H2 10–30 1 (cid:50)14.0 14.1 Holothuroidea Japanese trepang Apostichopus japonicus 14 May 99 H2 10–30 5 (cid:50)15.7(cid:54)0.4 10.2(cid:54)0.5 14 May 99 H4 10–30 1 (cid:50)14.9 11.7 Chordata Ascidiacea Styela plicata (Styelidae) 14 May 99 H1 10–30 1 (cid:50)18.2 13.4 By contrast, the average (cid:100)15N (8.2 (cid:54) 0.8‰) of the am- on 22 December 2000 was 9.8‰ in the size fraction of 15– phipods was 3.4‰ more depleted than the (cid:100)15N of the small 32 (cid:109)m at the sublittoral fringe where Bacillariophyceaewas decapods,beingasdepletedasthe(cid:100)15Nofprimaryproducers predominant,whilethe(cid:100)15Nwas3.8‰inthefractionof63– such as macroalgae and EOM (0.7–125 (cid:109)m) (Table 3, Fig. 125 (cid:109)m at the littoral fringe (high shore) where Cyanophy- 6). This difference of (cid:100)15N between small decapods andam- ceae was predominant (Takai unpubl. data). This suggests phipods may be related to the (cid:100)15N variation in microalgae that the amphipods may use the 15N-depleted organic matter inhabiting the shore. The (cid:100)15N of EOM collected at Sta. Z produced at the littoral fringe. The (cid:100)15N of the fish at Sta. Z (except the young black rockfish) ranged from 13.4 to 15.8‰ and was 1.8–4.2‰ more enriched than the average (cid:100)15N of 11.6 (cid:54) 0.2‰ in the decapods at Sta. Z (Fig. 6). This difference was consistent withthe(cid:100)15Nincreaseof3.4(cid:54)1.1‰pertrophiclevel(Min- agawa and Wada 1984), thus suggesting that a large portion ofthefishcollectedatSta.Zwouldbesecondaryconsumers in a food chain based on benthic primary producers. Like- wise the (cid:100)15N of 93% of the 89 fish collected in the deeper area of the central bay (Stas. H3 and H4) was distributed from14.0to16.0‰.Thissuggeststhatthefishinthedeeper area would also be mostly secondary consumers dependent on the benthic primary producers. Here it is necessary to pay attention to potential mixed feeding on prey from dif- ferent trophic positions. It appears that such complexity in Fig. 4. The local differences of (cid:100)15N in cardinal fish Apogon feeding links varied the (cid:100)15N values of the fish and as a lineatus (n (cid:53) 45) and young-of-the-year gurnard Lepidotrigla mi- consequence distributed the (cid:100)15N continuously in the range croptera (48–76 mm; n (cid:53) 26). from 14.0 to 16.0‰ (Tables 2, 4). Carbon sources for demersal fish 739 Fig. 6. The (cid:100)13C–(cid:100)15N food web diagram for Hiroshima Bay. Thefishweredividedintothreedistinctgroups(shadedareas);(A) fish dependent upon water column primary production, (B) migra- Fig.5. Therelationshipbetweenstandardlengthandstableiso- tors from the open sea, and (C) fish dependent on benthic primary tope ratios in Japanese anchovy Engraulis japonicus collected at production. The seasonal means of POM (0.7–125 (cid:109)m) in the sur- Stas. Y, H1, and H2 (n (cid:53) 24). (A) (cid:100)13C. (B) (cid:100)15N. face layer of area W from 14 May 1999 to 10 November 2000 (n (cid:53) 10) and EOM at the sublittoral fringe of Sta. Z from 10 August 1999 to 12 October 2000 (n (cid:53) 7), and the mean of SOM in the The maximum (cid:100)15N value among the fish was found in bay (n (cid:53) 4)are also shown.(a)Theyoung-of-the-yearblackrock- the cutlassfish of 17.5‰ at Sta. H4 (Table 4). Since this fish Sebastes inermis in June (Sta. Z; n (cid:53) 17). (b) Theotherblack value was 5.9‰ more enriched than that of the decapods at rockfish at Sta. Z (n (cid:53) 33). (c) All fish at Sta. Z except black Sta. Z (11.6 (cid:54) 0.2‰), it is evident that this carnivorousfish rockfish(n(cid:53)139).(d)JapaneseanchovyEngraulisjaponicus(cid:36)125 would be the typical tertiary consumer in the bay. mm (Sta. Y; n (cid:53) 4). (e) Japanese anchovy (cid:35)120 mm (Stas. H1, The (cid:100)15N values of the fish collected in the deeperareaof H2, and Y; n (cid:53) 20). (f) Gizzard shad Konosirus punctatus (Stas. the northern bay (Stas. H1 and H2) were slightly more en- H1–H4, Y; n (cid:53) 42). (g) Cutlassfish Trichiurusjaponicus(Sta.H4; riched than those of fish at Stas. H3 and H4, being distrib- n (cid:53) 1). (h) All fish in the deeper zone except the anchovies, the shad, and the cutlassfish (Stas. H1, H2, H3, H4, and Y; n (cid:53) 145). uted from 14.0 to 16.0‰ in 66% and over 16.0‰ in 30% (i) Amphipods (Sta. Z; n (cid:53) 14). (j) Isopod (Sta. Z; n (cid:53) 1). (k) of the fish (Table 4). Although this might suggest that fish Small decapods (Sta. Z; n (cid:53) 3). (l) Large decapods (Sta. H1; n (cid:53) in the north had higher trophic positions, it is more likely 2). (m) Mysids (Acanthomysis tenuicauda and Acanthomysis spp.; that the higher (cid:100)15N values were simply a reflection ofbase- Sta.H4;n(cid:53)1).(n)Copepods(Sta.H4;n(cid:53)1).(o)Japanesesquid line enrichments of POM (Table 1) and macroalgae (Takai Loliolusjaponica(Sta.H2,n(cid:53)2).(p)Theothercephalopod(Stas. et al. 2001) that were higher in the north than in the central H1 and H4, n (cid:53) 3). (q) Starfish Asterias amurensis (Sta. H2, n (cid:53) bay. This baseline variation was clearly reflectedinthe(cid:100)15N 1). (r) Japanese trepang Apostichopus japonicus (Stas. H2 andH4; values for the cardinal fish and the young-of-the-year gur- n(cid:53)6).(s)JapaneseoystersCrassostreagigas(Stas.Z,H3,andX; nard (Fig. 4). In aquatic ecosystems, 15N-enriched nitrogen n (cid:53) 20). (t) Egg cockles Fulvia mutica (Sta. H4, n (cid:53) 4). (u) As- from wastewater increases (cid:100)15N of the aquatic life (Mc- cidian Styela plicata (Sta. H1; n (cid:53) 1).Macroalgae(v–y)werecol- lectedatSta.ZbyTakaietal.(2001)andseagrassesZosteramarina Clelland et al. 1997). Since 82% of total nitrogen load from (z, z) were collected in the western Seto Inland Sea by Takai rivers that feed into Hiroshima Bay flows into the northern 1 2 (unpubl. data). (v) Phaeophyceae (n (cid:53) 30). (w) Rhodophyceaeex- area (Lee and Hoshika 2000), the (cid:100)15N increases in the pri- cept samples with special (cid:100)13C values of (cid:44)(cid:50)30‰ (n (cid:53) 20). (x) mary producers and fish were considered to reflect the en- Ulvophyceae (n (cid:53) 8). (y) Chlorophyceae (n (cid:53) 3). (z) A seagrass riched 15N signature of this wastewater. collectednearSta.H1on25Oct1999.(z)Seagrasse1scollectedin 2 We note that the fish analyzed in this study were mostly the embayment of the sea from 19 July to 10 October in 2000 (n adults. The clearly depleted(cid:100)13Cvaluesintheyoung-of-the- (cid:53) 7). The values are shown as mean values (cid:54)SD.