INTRODUCTION
Recently the importance of organisms to connect spatially isolated ecosystems has been increasingly recognized (Polis et al., Reference Polis, Anderson and Holt1997; Nakano & Murakami, Reference Nakano and Murakami2001; Hyndes & Lavery, Reference Hyndes and Lavery2005; Melville & Connolly, Reference Melville and Connolly2005; Schindler et al., Reference Schindler, Leavitt, Brock, Johnson and Quay2005). Many studies have used stable isotopes to estimate the food sources of benthic consumers in marine ecosystems (e.g. Wainright et al., Reference Wainright, Weinstein, Able and Currin2000; Kharlamenko et al., Reference Kharlamenko, Kiyashko, Imbs and Vyshkvartzev2001; Riera et al., Reference Riera, Stal and Nieuwenhuize2002; Doi et al., Reference Doi, Matsumasa, Toya, Satoh, Mizota, Maki and Kikuchi2005; Melville & Connolly, Reference Melville and Connolly2005). Estuaries, one of the most productive types of ecosystem (Costanza et al., Reference Costanza, d'Arge, Groot, Farber, Grasso, Hannon, Limburg, Naeem, O'Nell, Paruelo, Raskin, Sutton and Belt1997), are characterized by inputs of terrestrial and marine organic matter in addition to autochthonous production by macroalgae, microalgae, and seagrasses.
Macroalgal and seagrass communities are widely distributed in marine and brackish shallow water and have high productivity in those ecosystems (McRoy & McMillan, Reference McRoy, McMillan, McRoy and Helfferich1977; Zieman & Wetzel, Reference Zieman, Wetzel, Phillips and McRoy1980; Duarte, Reference Duarte1989). The primary production at marine costal areas is often transported to intertidal flats by waves and currents and beached on the flats, and this sometimes supports relatively high biomass of intertidal gastropods such as limpets (Bustamante et al., Reference Bustamante, Branch and Eekhout1995; Mann, Reference Mann2000; Wainright et al., Reference Wainright, Weinstein, Able and Currin2000). At our two study sites, one sub-tropical and the other temperate (26°N and 38°N, respectively), high numbers of Batillariid gastropods were observed on green macroalgae (Ulva and Enteromorpha) and on the leaves of a seagrass Zostera marina (Figure 1). Therefore, we hypothesized that the macroalgae and seagrasses are significant food sources for Batillariid gastropods on intertidal flats in both sub-tropical and temperate regions, which have a different composition of snail species. Since little is known about the importance of macroalgae and seagrass to gastropod foraging, and since numbers of gastropod snails are decreasing in Japan due to habitat loss and degradation by human activity, testing the above hypothesis would be an important contribution to conservation of local gastropod populations (Kimura & Kimura, Reference Kimura and Kimura1999; Kimura, Reference Kimura2005).
Fig. 1. (A) Sampling site at Station A in sub-tropical region with macroalga Enteromorpha sp. Photograph by M. Matsumasa; (B) sampling site at Station B in temperate region with Ulva pertusa and seagrass Zostera marina. Photograph by M. Fujikawa; (C) gastropods (Batillaria multiformis and Batillaria cumingi) were observed on macroalgae Ulva pertusa and seagrass Zostera marina at Station B. Photograph by M. Fujikawa.
Since a few elements can routinely be used for stable isotope analysis, food web studies using isotope data from n stable isotope values (typically two) have often restricted their studies to n+1 organic matter sources by including only those sources either assumed to be most important or shown to be through other types of analyses. Phillips & Gregg (Reference Phillips and Gregg2003) developed isotope mixing model software (IsoSource) that is designed for situations in which n isotopes are being used and more than n+1 sources are likely to be contributing to a mixture. IsoSource mixing model can calculate feasible ranges of source contributions using stable isotope data. We performed comparisons of food sources among different gastropod species at the sub-tropical and the temperate tidal flats (26°N and 38°N), by the isotope mixing models using carbon and nitrogen stable isotopes.
MATERIALS AND METHODS
Study sites
The sub-tropical site (Station A) is located on a sheltered intertidal flat with a mangrove forest (mainly Kandelia candel and Bruguiera gymnorrhiza, approximately 300-m away from flat; see Figure 1) in Ohura Bay, Okinawa Island (26°33′N 128°02′E). In March 2004, a green macroalgae Enteromorpha sp., which would be transported, covered the tidal flat (Figure 1). The temperate site (Station B ) is located on an open intertidal flat with salt marsh plants (mainly Carex scabrifolia and Phragmites australis) in Matsushima Bay, north-east Honshu Island (38°22′N 141°05′E). There was a seagrass bed (mainly Zostera marina) in the coastal zone around Station B, and the tidal flat was covered with beached leaves of the seagrass Z. marina and the macroalgae Ulva pertusa (Figure 1). The vegetation of seagrass was not found on the tidal flat of Station B. The sediment differed between the two sites: sand and cobbles at Station A, silt and sand at Station B.
Sample collection and preparation for isotopic analyses
The monthly mean air temperature was 16.3°C in March 2004, as recorded at the Japan meteorological observation site nearest to Station A (26°35′N 127°57′E), and 14.5°C in October 2003 at the site nearest to Station B (38°15′N 140°54′E) (data from Japan Meteorological Agency). The choice of sampling periods was made so as to minimize the difference in mean air temperature between the two sites, hence samples were collected at Stations A and B in March 2004 and October 2003, respectively. During the sampling periods, daily tidal amplitude ranged from 82–86 cm at the tidal observation site nearest to Station A (Itoman: 38°18′N 121°70′E), and from 89–96 cm at the site nearest to Station B (Ishihama: 26°80′N 127°40′E).
Three or four replicates of each sample were collected for isotope measurement. From the intertidal areas we collected the dominant adult Batillariidae species (shell lengths > 20 mm): Cerithidea (Cerithideopsilla) cingulata, Cerithidea (Cerithideopsilla) djadjariensis, and Batillaria flectosiphonata at Station A, and Cerithidea (Cerithideopsilla) rhizophorarum, Batillaria multiformis, and Batillaria cumingi at Station B. The stomach contents analysis for C. cingulata, C. djadjariensis, C. rhizophorarum and B. multiformis showed that these species fed on micro and macroalgae on the tidal flat (Maki et al., Reference Maki, Ohtaki and Tomiyama2002), thus probably the Batillariidae species had a similar trophic niche. In addition, at Station A we collected C. djadjariensis and B. flectosiphonata from two supralittoral sites to compare the spatial variations in the isotope values of the species. The distance between the supralittoral sites and the intertidal site was approximately 150 m. The substrate compositions of supralittoral sites were similar to that of Station A. These gastropods were the dominant species of the whole benthic community at each station (Motonaga, Reference Motonaga1977; Meziane & Tsuchiya, Reference Meziane and Tsuchiya2000; Fujikawa, M., unpublished data). We collected the gastropod feet, and foot lipids were removed using a chloroform–methanol mixture (2:1 by weight). We analysed the isotopic signature of each individual.
Sediment samples containing sediment organic matter (SOM) were collected from depths of 0–0.5 cm below the sediment–water interface at each station. The green macroalgae Enteromorpha sp. and Ulva pertusa were collected on the tidal flat of Stations A and B, respectively, while the leaves of the dominant terrestrial plants Kandelia candel and Carex scabrifolia were collected at Stations A and B, respectively. In addition, at Station B, we collected the beached leaves of seagrass Zostera marina from the offshore area to the tidal flat, but we did not observe any beached seagrass leaves at Station A (Figure 1). Macroalgae and seagrass were washed with dissolved water, and any attached microorganisms and detritus (i.e. benthic diatoms, bacteria, organic matter trapped within mucus secreted by benthos, etc.) were gently scraped off using a tooth brush. Most of the attached microorganisms and detritus were removed by this method (Aikins & Kikuchi, Reference Aikins and Kikuchi2002). Crushed and homogenized macroalgae, seagrass, and SOM samples were put in a glass vial and acidified with 1 mol l−1 HCl to remove carbonates before the isotope analysis.
Benthic diatoms were collected from the sediment at Station B, making use of their phototactic movement (Couch, Reference Couch1989; Doi et al., Reference Doi, Kikuchi, Hino, Itoh, Takagi and Shikano2003a, Reference Doi, Kikuchi, Hino, Itoh, Takagi and Shikanob). Sediment samples from a depth of 1 cm were collected and spread in a 1-cm-thin layer on Petri dishes. For each sample, nylon net (75-µm mesh) was placed on the sediment surface and a 2-mm layer of 25- to 65-µm silica powder (precombusted at 500ºC for 2 hours) was spread on the net. The Petri dishes were illuminated for 24 hours by continuous spraying of filtered river water to retain moisture in the silica powder. After 24 hours, the powder was scraped off and mixed with filtered water to suspend the benthic diatoms, which were then decanted into glass vials and freeze-dried. Unfortunately, at Station A we could not collect benthic diatoms because it was difficult to extract them from the muddy sediment using the 75-µm mesh net. This is one limitation of this study, and the reason why the isotope mixing model was performed only for Station B, as described below. Other potential food sources such as other macroalgal species were not observed at Station A (Doi, H., personal observation).
Stable isotope analysis
All the samples were freeze-dried and preserved in a freezer at −20°C prior to isotope ratio and elemental analyses. For nitrogen isotope ratios, the samples were measured without acid-treatment. The carbon and nitrogen isotope ratios of the samples were measured with a mass spectrometer (DELTA plus, Finnigan Mat), and the results are reported using delta notation obtained by δ13C or δ15N = (Rsample/Rstandard−1) · 1000 (‰), where R is the 13C/12C or 15N/14N ratio for δ13C or δ15N, respectively. As standards, PDB was used for δ13C and air nitrogen was used for δ15N. The analytical error was±0.2‰ for both δ13C and δ15N.
Isotope mixing model and statistical analyses
To determine which of the potential food sources were assimilated by the snails, we estimated the feasible contributions for each food source by isotope mixing models using IsoSource version 1.3.1 software (Phillips & Gregg, Reference Phillips and Gregg2003). Essentially, the model iteratively creates all possible combinations of source proportions (with each combination equalling 100%) at preset increments (2% in this study) to create a set of predicted mixtures of sources (see Phillips & Gregg, Reference Phillips and Gregg2003, in detail). For the model, the potential food sources selected were SOM, macroalgae and terrestrial and mangrove plant litter for Station A and SOM, macroalgae, seagrass, benthic microalgae and terrestrial plant litter for Station B. Tolerance was initially set at 0.1‰; if mixture isotope values were out of bounds (i.e. the snails at Station A), we incrementally increased the tolerance value by 0.1‰ up to a maximum of 0.8‰. The carbon and nitrogen isotope enrichments of the snails were +0.8 and +2.3‰ (France & Peters, Reference France and Peters1997 for carbon; McCutchan et al., Reference McCutchan, Lewis, Kendall and McGrath2003 for nitrogen).
To compare the isotope values of the potential food sources and snails, we performed one-way analysis of variance (ANOVA) and Tukey multiple comparison. In addition, at Station A we compared the isotope values of the snails at mid-intertidal flats (Station A) and supralittoral flats (Stations A-SF1 and A-SF2) using repeated-measures two-way ANOVA.
RESULTS
Isotopic signatures of the samples
Carbon and nitrogen isotope signatures of the potential food sources varied among the sources at Stations A and B (Table 1; Figure 2). At Station A the mean carbon and nitrogen isotope values of the snails ranged from −14.7 to −13.9‰ and from 11.3 to 12.8‰, respectively. Nitrogen isotope values of the snails at Station A were significantly different among species (one-way ANOVA; Table 1), and those of Batillaria flectosiphonata were significantly lower than those of other species (Table 1; Tukey multiple comparison, P < 0.05). The carbon isotope values were not significantly different between species (Table 1). Comparison of the isotope values of Cerithidea (Cerithideopsilla) djadjariensis and B. flectosiphonata between the intertidal and supralittoral sites at Station A (Table 1) revealed that carbon and nitrogen isotope values were not significantly different between the sites (ANOVA, F = 2.95 and 0.66, P = 0.08 and 0.53, respectively). Thus, the spatial variations of the isotope values of the snails were small over a small horizontal scale at Station A.
Fig. 2. Carbon and nitrogen stable isotope plots of the snails and their potential food sources at Station A (upper) and Station B (lower). Each symbol shows a mean±1 SE.
Table 1. Carbon and nitrogen stable isotope values of samples at Stations A and B (Mean±SE, N = 3 or 4). SOM, mean sediment organic matter. A-SF1 and 2 are two sites on supralittoral flats at Station A. F = mean F-value of ANOVA (**, * and non-asterisk indicate P < 0.01, < 0.05 and > 0.05, respectively). Tukey multiple comparison results show that different letters indicate significant differences (P < 0.05).
At Station B, the mean carbon and nitrogen isotope values of the snails ranged from −12.7 to −12.5‰ and from 10.8 to 11.6‰, respectively. Nitrogen isotope values of the snails at Station B were significantly different between species (F = 7.58, P = 0.02), and those of C. rhizophorarum were significantly lower than those of B. multiformis (P < 0.05). However, the carbon isotope values were not significantly different (F = 3.65, P = 0.09). Carbon isotope ratios of benthic microalgae were −23.0± 0.2‰. These values are similar to the values (−23.0 ± 0.5‰) obtained at the tidal flats studied by Guest et al. (Reference Guest, Connolly and Loneragan2004).
Contribution of the food sources
Using the IsoSource mixing model, we estimated the contribution of the potential food sources to the snail biomass (Table 2). At Station B, food sources for three snails were estimated to be mainly macroalgae (50–56%) and seagrass (39–45%). The contributions of terrestrial plants, SOM and benthic microalgae were smaller than those of macroalgae and seagrass. At Station A, we could not perform IsoSource mixing models to the contributions for snails, since isotope values of the snails were higher isotope values from any potential food sources. This suggested that the snails fed mainly on macroalgae at Station A. We could not detect the remarkable differences in food sources for snail species, since their carbon isotope values were not significantly different, although the nitrogen values were slightly different.
Table 2. Contribution of the potential food sources to snails using IsoSource mixing model for Station B. POM, mean particulate organic matter.
DISCUSSION
At both the sub-tropical and the temperate site (Stations A and B), our results showed the significant contribution of macroalgae and/or seagrass to snails as food sources. In this study, we also observed that snails attached the macroalgal mats of Enteromorpha at Station A, and Batillaria multiformis and B. cumingi attached both the macroalgae and the leaves of seagrass at Station B. Meziane & Tsuchiya (Reference Meziane and Tsuchiya2000), using fatty acid analysis, suggested that Batillaria zonalis, always observed on a macroalgal mat, mainly fed on green macroalgae. At Station B, we did not observe Cerithidea (Cerithideopsilla) rhizophorarum colonizing the macroalgae or the leaves of seagrasses. However, the main food sources for C. rhizophorarum were estimated to be also macroalgae and seagrass, although its nitrogen isotope value was slightly lower than that of B. cumingi.
The results of our mixing model showed the high contribution (39–45%) of seagrass leaves to snails on the tidal flats at the temperate site. Live tissues of marine vascular plants are of limited use as food for marine animals; therefore, other primary producers or the detritus food chain support secondary production in these communities (e.g. Mann, Reference Mann1972; Fenchel, Reference Fenchel, McRoy and Helfferich1977; Kharlamenko et al., Reference Kharlamenko, Kiyashko, Imbs and Vyshkvartzev2001). By observing feeding, Whitlatch & Obrebski (Reference Whitlatch and Obrebski1980) reported that the gastropod Cerithidea californica occasionally feeds on large vascular plant fragments. Also, Kharlamenko et al. (Reference Kharlamenko, Kiyashko, Imbs and Vyshkvartzev2001) suggested that the gastropod B. cumingi derived on average 50% of its carbon from Z. marine in the subtidal seagrass bed. Thus, seagrass species would play an important role as food sources on tidal flats as well as on the seagrass bed of the costal zone.
Increasing attention has focused on benthic microalgae as a food source for the many benthic invertebrates on the tidal flats (e.g. Currin et al., Reference Currin, Newell and Paerl1995; Doi et al., Reference Doi, Matsumasa, Toya, Satoh, Mizota, Maki and Kikuchi2005; Kanaya et al., Reference Kanaya, Nobata, Toya and Kikuchi2005). For Gastropoda, Page (Reference Page1997) used stable isotopes to show that Cerithidea californica, inhabiting tidal flats, feeds principally on benthic microalgae. In this study, however, the contributions of benthic microalgae to the snails were very low. Using fatty acid analysis, Meziane & Tsuchiya (Reference Meziane and Tsuchiya2000) suggested that the contribution of diatoms to C. cingulata was weaker than that of green macroalgae and bacteria at Ohura Bay (Station A in this study). Doi et al. (Reference Doi, Kikuchi, Takagi and Shikano2006) also suggested that deposit-feeding chironomid larvae assimilated green algae more than benthic diatoms. The snails in this study may assimilate the green macroalgae more easily than benthic microalgae (mainly diatoms), and feed mainly on the macroalgae in the algal patches on the tidal flat. In addition, recent studies have suggested that the contribution of terrestrial organic matter is lower than that of other sources in food webs on tidal flats (e.g. Kurata et al., Reference Kurata, Minami and Kikuchi2001; Doi et al., Reference Doi, Matsumasa, Toya, Satoh, Mizota, Maki and Kikuchi2005). Our mixing-model results revealed the lower contribution of mangrove and salt marsh plants to the snails and supported the low contribution of terrestrial organic matter in their systems, although the vegetation of the sub-tropical and temperate sites were quite different.
In conclusion, we showed that macroalgae and seagrass made a greater contribution to snail biomass than the other potential food sources, and thus, that macroalgae and seagrass play an important role as food sources not only at their growing sites but also on tidal flats. We should acknowledge the limitation of this study. This collection was conducted in a season, however, the biomass of macroalgae and seagrass may vary seasonally. Thus, the food source composition of the gastropods would be changed along with biomass of macroalgae and seagrass. Future studies are needed to reveal the seasonal dynamics of the food sources for gastropods. To conserve local gastropod populations, we should consider the connection between tidal flats and coastal marine ecosystems in ensuring their supply of food.
ACKNOWLEDGEMENTS
We are grateful to Dr K. Itoh, Faculty of Agriculture, Tohoku University for providing analytical facilities for stable isotope analysis. This work is supported in part by the Japan Society for the Promotion of Science (H. Doi.) and a Grant-in Aid for Scientific Research (C) (No. 14540584) (M. Matsumasa.)
