Feeding Ecology And Food Relations Of The Main Fishes In The Beibu Gulf, South China Sea

Marine fish feeding ecology and food relationships are the main contents of marine ecotrophic dynamics studies. The Beibu Gulf of the South China Sea, a tropical and subtropical semi-closed bay, is the traditional fishing ground of China and Vietnam. Studies on the fish ecology and foraging associations in this sea area are of significant academic and practical value. Based on monthly sampling in the main fishing ports of the Beibu Gulf from October 2008 to September 2009 and on the fishery-dependent trawl survey in February 2010,stomach contents analysis and stable isotope analysis were used to study the prey category,trophic position and ontogenetic feeding habit shifts of the main fishes in the Beibu Gulf,South China Sea. The main results are as follows:(1) Trichiurus lepturus mainly feeds on 43 species of prey. Among them, Bregmaceros rarisquamosu was the principal prey component of the diet, accounting for 37.99% of the contents on the Index of Relative Importance. Small pelagic fish, Decapterus maruadsi, and crustacean, Acetes chinensis, with the percentages of 16.42% and 10.03%, respectively, were also relatively important in the diet of this fish, whereas the other 40 prey species, including pisces, cephalopoda, benthic crustacean and zooplankton, were only present in lower percentages.Since Decapterus maruadsi and Anchoviella heteroloba appeared throughout the Beibu Gulf area year-round, they could serve as index species for the largehead hairtail migration and fishing ground distribution. The feeding intensity and stomach fullness indexes differed significantly from month to month (P<0.001). However, niche width, measured by Shannon-Weiner diversity index H′, indicated no significant seasonal differences (P>0.05). H′reached its peak in autumn and had an average value of 1.97. Cluster analysis, measured with Bray-Curties similarity index, revealed a clear diet shift at preanal length 190mm, i.e. at 50% of maturity length. It was concluded that ontogenetic development was synchronized with the changes in feeding habits from small benthic, zooplankton and pelagic fishes for juvenile largehead hairtail to predominantly larger fish and cephalopoda for adult hairtail. (2) Stomachs of 666 hairtail, Trichiurus margarites Li 1992, ranging in size from 121 mm to 561 mm in preanal length (PL), were sampled monthly in the Beibu Gulf of the South China Sea, and analysis of their stomach contents showed Decapterus maruadsi to be the primary prey components of the diet, averaging 23% by number, 34% by weight and 49% by Index of Relative Importance (IRI). Small benthic fish, Bregmaceros rarisquamosus and B. nectabanus, and crustacean, Acetes chinensis, with IRI percentages of 8%, 4% and 3%, respectively, were also relatively essential in the diet of this fish, whereas the other 30 species of prey were only present in comparatively lower percentages. The research findings revealed that the mean repletion index varied considerably by month, while trophic diversity, calculated by Brillouin index Hz, proved to be relatively consistent throughout the year. Mean repletion index increased in accordance with the gonad maturity stages of hairtail. On the other hand, the proportion of empty stomachs decreased with the stages, being the highest at Stage I with 60.9% and the lowest at StageⅥwith 41.6%. The 14 PL classes demonstrated significant differences with respect to prey weight and numbers. The tendency of mean prey number and weight increased with the body size. Using PRIMER 5.2 software, the cluster analysis, based on IRI values for the above PL classes, showed that the feeding habits of hairtail were relatively heterogeneous among different sizes. In conclusion, the diet of hairtail, T. margarites varied significantly by month and ontogenetic development. These findings may later assist in the development and improvement of marine ecosystem modeling for the Beibu Gulf, South China Sea.(3) As each curve of trophic diversity calculated with the Brillouin index reached an asymptote, the number of stomachs examined in this study was sufficient to describe the monthly and ontogenetic diet shifts for Trichiurus lepturus and T. margarites. T. lepturus is a general omnivore, feeding mainly on Bregmaceros rarisquamosu, Decapterus maruadsi and Acetes chinensis, while T. margarites preys primarily on D. maruadsi, B. rarisquamosus, B. nectabanus and A. chinensis. Both species showed distinct ontogenetic diet shifts in the studied periods. Even though the two examined hairtail feed on similar resources in the same habitat, the diet overlap was relatively low. The results in this study indicated that hairtail played an important role in controlling small pelagic and benthic fish population in the Beibu Gulf area. This trophic correlation may later assist in the establishment of marine ecosystem modeling for the Beibu Gulf of the South China Sea.(4) Stomachs of 749 greater lizardfish, ranging in size from 124 mm to 375 mm in standard length (SL), were sampled monthly in the Beibu Gulf of the South China Sea, and analysis of their stomach contents showed Decapterus maruadsi to be the primary prey components of the diet, averaging 9.7% by number, 55.5% by weight and 34.9% by percentage of Index of Relative Importance (IRI). Small benthic fish, Bregmaceros rarisquamosus and pelagic fish, Thryssa setirostris and Leiognathus riviulatus, with IRI percentages of 26.6%, 4.3% and 2.5%, respectively were also relatively essential in the diet of this fish, whereas the other species of prey were only present in comparatively lower percentages. The research findings revealed that the mean repletion index and mean maturity coefficient varied considerably by month, while trophic diversity, calculated by Brillouin index Hz, proved to be relatively consistent throughout the year. Mean repletion index (RI) was highest (RI = 7.61%) at the stage ofⅥ-Ⅱindicating that this group of greater lizardfish reverted from StageⅥto StageⅡincreased their food input to compensate for energy loss during reproduction. On the other hand, the proportion of empty stomachs reached its peak with 61.5% while mean RI reached the second highest position with 5.98% at StageⅤ, indicating that during the reproduction periods from March to April, and in June and November, greater lizardfish cease feeding until digestion of the prior meal was completed or near completion. The 12 SL classes demonstrated significant differences with respect to prey weight but showed relative consistency in prey numbers. The tendency of mean prey weight increased with the body size following the improvement of the feeding organs and swimming ability of greater lizardfish. Using PRIMER 5.2 software, the cluster analysis, based on IRI values for the above SL classes, showed that the feeding habits of greater lizardfish were generally homogeneous among different sizes, and particularly among the 136~255mm SL classes, with similarities of more than 60%. In conclusion, the diet of greater lizardfish, Saurida tumbil varied significantly by month and ontogenetic development. These findings may later assist in the development and improvement of marine ecosystem modeling for the Beibu Gulf, South China Sea.(5) A total of 645 (344 stomachs with contents) dorab wolf-herring, ranging in size from 193 mm to 782 mm in fork length, were randomly sampled through trawl and gillnet commercial fishing. The results showed that the dorab wolf-herring of the Beibu Gulf was omnivorous and predated not only on small pelagic fish but also on benthic fish, shrimps, crabs and squids. Pisces were the primary prey components of the diet with a percentage of Index of Relative Importance (%IRI) of 99.7%. Among these, 24 species were identified. Genera such as Bregmaceros sp., Thryssa sp., Sardinella sp. and Decapterus sp. were dominant in the dorab wolf-herring food constituent. The small fish, Stolephorus zollingeri, Bregmaceros rarisquamosus, Sardinella jussieu, Decapterus maruadsi, Stolephorus commersoni and Caranx mate, were also relatively important in the diet of this fish, with the weight percentages of 10.27%, 9.36%, 4.72%, 4.11%, 2.77% and 2.73%, respectively, whereas the other species prey were only present in lower percentages. The research findings revealed that the mean repletion index and vacuity coefficient varied considerably by month, while trophic diversity, calculated by Brillouin index Hz, proved to be relatively consistent throughout the year. The spawn peak, from May to August, was recognized by the female gonad somatic index (GSI). Furthermore, the peak periods were consistent with the surface water temperature,20m and 30m layers provided by France CATSAT fisheries remote sensing system. It can be safely concluded that spawning and reproduction of dorab wolf-herring, one of the pelagic fishes, correlate with the sea surface temperature. Mean repletion index (RI) was highest (RI = 2.50%) at the stage ofⅤindicating that this group of dorab wolf-herring maintained strong feeding intensity during the reproduction periods. Meanwhile the vacuity coefficient was almost lowest (VC = 0.75%) at the stage ofⅥ-Ⅱindicating that the fish reverting from StageⅥto StageⅡincreased their food input to compensate for energy loss during reproduction. The 12 fork -length classes demonstrated significant differences with respect to prey weight and prey numbers. The tendency of mean prey weight and number increased with the body size following the improvement of the feeding organs and swimming ability of dorab wolf-herring. Using PRIMER 5.2 software, the cluster analysis, based on %IRI values for the above fork-length classes, showed that the feeding habits of dorab wolf-herring could be classified into three groups: under 380mm, between 381 to 530mm, and above 531mm. The absence of a significant linear correlation between the fork length andδ13C lead to the conclusion that dorab wolf-herring were feeding all the water layers during the major part of their biological life cycle. With an average value of 3.4, the trophic level of dorab wolf-herring calculated byδ15N varied not significantly from the fork length. It can be concluded that stable isotope analysis will play an import role in evaluating the fish feeding ecology and in constructing the marine food web.(6) The annulus of Pennahia pawak in the scales were of the vacant type, which indicated that they formed once a year, during September– November by monthly changes in the marginal growth increment (MGI )of the scales, just after the spawning had occurred from April to August. A maximum likelihood estimation procedure was used to fit the von Bertalanffy, Logistic, Gompertz and generalized von Bertalanffy growth functions to the length-at-age data. ARSS indicated that there were no significant differences in growth between sexes in the four growth models (P>0.05), and the Logistic growth function was selected as the most appropriate growth model according to Akaike's Information Criterion (AIC), AIC differences( ) and Bayesian information criterion(BIC.The estimated parameters for Logistic growth function were LΔi∞= 220.32 mm,K = 0.58,t0 = 0.905. P. pawak showed rapid growth before the 3rd year, after which the growth tended to slow down; therefore, the starting age of 3 was proposed the best stage of exploitation and utilization for this resource. Pennahia pawak preyed mainly on benthic organisms. Analysis of their stomach contents showed that by Index of Relative Importance (IRI), fish and crustacean were the primary prey components of the diet, accounting for 57.0% and 43.0%, respectively. Bregmaceros rarisquamosus, B. nectabanus and Acetes chinensis were the dominant species in the 27 prey species, making up 6.55%, 6.80% and 5.27% respectively. Solenocera crassicorni and Thryssa setirostris also appeared in all seasons. The feeding intensity and repletion index differed significantly according to seasons (p<0.001), while food diversity had no considerably shifts during the whole year (p>0.05). Cluster analysis based on %IRI values showed that the feeding habits of P. pawak shifted at the standard length of 130mm, which is a little bigger than the 50% sex maturity length (115mm).(7) Small pelagic fish,Decapterus maruadsi,Sardinella jussieu,Stolephorus heteroloba,benthic Bregmaceros rarisquamosus and cephalopoda,Loligo chinensis, were also relatively important in the diet of this fish,with the weight percentages of 26.19%,10.21%,9.94%,7.20% and 6.07%,respectively,whereas the other species prey were only present in lower percentages. Theδ13C signatures increased negative (depleted) values for pelagic species to more enriched values for benthic species. As expecte, D. maruadsi displayed the lowest value (-17.830‰) while benthic Leiognathus lineolatus displayed the highest value (-14.925‰). The absence of a significant linear correlation between the preanal length andδ13C leads to the conclusion that T. lepturus are feeding all the water layers during the major part of their biological life cycle. Both the average trophic level and its annual mean value were calculated by prey weight percentages andδ15N are 3.7. The trophic level of T. lepturus varied not significantly with the preanal length,despite the fact that in the same sea area T. margarites and T. minor increased notably. Two samples of dehydration undertaken prior to the analytical determination of stable isotope ratios were tested and resulted in the observation that vacuum freeze-drying and drying in an oven at constant temperature showed significant variance inδ15N but not inδ13C. Undoubtedly,SIA will play an import role in evaluating the fish feeding habits and trophic positions. At the same time,the following three key points should be taken into account: (a) SIA must be combined with the stomach contents analysis; (b) an appropriate baseline organism has to be selected and its exact trophic position determined; (c) attention needs to be paid to the uncertainty of SIA application in the ontogenetic diet shifts and trophic position variations.