The sandstone reefs located on the southern coast of Pernambuco state are marine outcrops forming ridges parallel to the coast that often emerge during low tide. These reefs consist mainly of quartz and bioclasts, primarily originating from shellfish and algae (Sousa and Cocentino, 2004). Many of these reefs represent a key consolidated substratum for the attachment of benthic algae and a diverse associated fauna, therefore representing an ecosystem with significant biological interact-tions (Simões et al., 2009) that forms the base of the local food webs.

The Tricolia genus has a wide geographical distribution and is found in North Carolina (USA) to the coast of Brazil. The species Tricolia affinis is a prosobranchiate gastropod of the family Tricoliidae found at coastal locations in Florida, the Caribbean, Colombia, Venezuela and Brazil, exhibiting sizes varying between 0.2 and 4.0 mm (Rios, 1994). On the Brazilian coast, this genus is found from Santa Catarina State to the northeastern region, including the Fernando de Noronha archipelago (Nogueira et al., 2003). According to Braga (1983), T. affinis is an abundant species in the studied area (Piedade beach). This gastropod is commonly found in estuaries and on rocky shores, coral reefs, unconsolidated substrata and seagrass banks, often associated with algae.

T. affinis is an herbivorous species that is most frequently found in the subtidal zone, where there are higher nutrient concentrations. This species feeds on microscopic algae and bacteria (periphyton) on the surface of debris-covered macroalgae (Marcus and Marcus, 1960). According to Ourives et al. (2011), T. affinis is also an abundant component of the estuarine fauna of Rio Cachoeira, Bahia State, Brazil, where its distribution was found to be positively correlated with salinity. The aim of the present study was to assess the temporal and spatial variation of T. affinis, associated with a Gracilaria bank, in the intertidal zone of the reef at Piedade beach since, this alga has been commercially used on the Brazilian coast, for agar production, and its abundance was found to be reduced in several locations along the northeastern coast. While in other areas, it is considered as an invasive species (Fukunaga et al., 2014; Zi-Min and Lopez-Bautista, 2014).

Two areas were sampled: a tidal pool where the algae were always immersed and more likely subjected to water transport, the reef flat algae where they are better attached to the substrate and subjected to less water transport. Therefore, the present study was based on the hypotheses that a correlation exists between the dominant algal biomass, abiotic factors and the abundance of T. affinis at the sampling sites and that spatial and temporal differences exists in the mollusk and algae abundance.

Laboratory and statistical analyses

The mollusks were separated on a plastic tray using a stereomicroscope, metallic tweezers and Petri dishes. Then, they were fixed in 70% ethanol for subsequent counting. Only T. affinis samples in good condition and with an intact operculum were included in the counts, with empty shells being excluded. The dominant algae were identified taxonomically at the genus and species level. Algae were transferred to laminated paper bags and dried at 60°C for 72 h to assess their biomass. The samples were then weighed on an analytical scale, and the algal biomass values were expressed per m². Bivariate and univariate analyses of variance (ANOVAs), in addition to Tukey’s tests, were used to assess the seasonal variation of algae at the reef flat and tide pool after logarithmic transformation of the biomass data.

Two-way ANOVAs were performed to assess the spatiotemporal variability of the T. affinis density in the samples, following confirmation of the normal distribution of the data and square root (x) + 0.5 transformation. Subsequently, one-way ANOVAs were performed for each sampling station individually, in addition to Tukey’s test. Canonical correspodence analyses were performed with the mean algal biomass and the mean density of T. affinis with the abiotic variables (water temperature, salinity, pH and total nitrogen) separately for the two sampling sites. The Pearson correlation coefficients between the dominant algal biomass and abundance of T. affinis were assessed. The statistical programs R (R Core Team, 2014) and Biostat 5.8.4 were used for these analyses.

A two-way ANOVA showed that there was significant temporal variation in the T. affinis density, but no significant differences were found in gastropod abundance between the sampling sites. The temporal variation of Tricolia was significant in the reef flat samples, significant differences between the sampling months were found (p < 0.01) (Figure 1). The mean gastropod density on the reef flat differed between April 2009 and June, August and October 2009. A significant difference was also found between October 2009 and March 2010, when the mean density of Tricolia was approximately 850 individuals per m², demonstrating a population recovery at the end of the dry season.

The  Rhodophyta  G.  caudata,  Hypnea  (Hypnea musciformis, Hypnea spinella and Hypnea cenomice) and the Chlorophyta Bryopsis (Bryopsis plumosa and Bryopsis pennata) were the dominant algae in terms of biomass (Figures 2, 3 and 4). These algal genera were found on all of the sampling dates at both the tide pool and reef flat. Their mean dry weights were subjected to statistical analyses in combination with the density of T. affinis. Ulva spp. was also found during all of the sampling campaigns on Piedade reef, but at a lower biomass than Gracilaria, Hypnea and Bryopsis and was therefore not included in the statistical analyses. The macroalgae showed seasonal variation, reflected by a lower biomass of G.caudata in rainy season samples (p < 0.05). However, the biomass of Hypnea showed no significant differences between the sampling months at either the tide pool or reef flat (p > 0.05) but was different between sites. No significant differences in Bryopsis biomass were found on the reef flat between the sampling dates (p > 0.05). In turn, Bryopsis showed a peak biomass in December 2009, which was significantly higher (p < 0.01) than the mean biomasses detected in the other sampling months at the tide pool, although no differences in biomass were found between sampling sites. A two-way ANOVA demonstrated a higher abundance of G. caudata at the reef flat than in the pool.

The T. affinis density was positively correlated with the Gracilaria biomass for the samples collected at both the reef flat (r = 0.91, p < 0.01) and tide pool (r = 0.88, p < 0.05). The salinity in the study area was low (15.26 – 20.45‰) during the rainy season, which may have been a limitation for T. affinis populations. The canonical correspondence analysis for the reef pool showed the correlation between total nitrogen and Hypnea biomass and Tricolia density close to Gracilaria biomass (Figure 5) where the mollusk was less variable, in the same site Gracilaria is opposite to Bryopsis. In the reef flat, salinity is close to Gracilaria which is opposite of Hypnea and Bryopsis (Figure 6).

The opposite canonical correspondence analyses ordering of algal genera indicates competition between functional groups (Gao et al., 2014), evidenced also by the analyses of variance since Hypnea spp. was more abundant at the tidal pool whereas G.caudata biomass was greater at the reef flat. A two-way ANOVA showed that there was temporal variation in the T. affinis density; however, there was no significant difference in the gastropod’s abundance between the two sampling sites. We can therefore conclude that the spatial distribution of T. affinis was proportional at the two sampling sites. Therefore, a correlation between T. affinis density and G. caudata biomass was found, at the reef flat and pool, possibly related to the alga’s higher structural complexity or the provision of greater protection for recruitment and from predation and ocean currents, in addition to being a reliable source of food. Gracilaria exhibited a lower biomass in the samples from the rainy period, directly affecting the number of sampled Tricolia. Gracilaria’s stem complexity confirms the correlation of T. affinis with its structural composition, which may be related to the retention of sediments that provide a greater supply of nutrients and protection against predation and hydro-dynamic counterflows for the mollusk (Chemello and Milazzo, 2002). Pereira and Jacobucci (2008) suggested that although T.affinis has been registered as a diet staple for fishes and crabs and an important link in the food chain of aquatic environments, it may not play a key role in the diets of marine species due to its low nutritional quality and calcareous shell (Dubiski-Silva and Setuko, 2008). The study area is a highly impacted region, constant trampling by bathers, which may be one reason for the success of the gastropod and relationship with Gracilaria. Invasive Gracilaria salicornia in the coast of Hawaii improved the density of epifauna and detritivores in comparison with Padina thivyi, while the invasive Acanthophora spicifera did not cause any epifauna improvement (Fukunaga et al., 2014).

The density of this gastropod was positively correlated with the biomass of G. caudata which showed the highest dry weight values at the sampling sites and presented a significantly higher biomass at the reef flat than the pool, where the temporal variation in Tricolia was less evident. Along with Tricolia, G. caudata showed seasonal varia-tion between the dry and rainy periods, with a higher biomass found in dry periods, when there was also a higher temperature and salinity (Silva et al. 1987; Marinho-Soriano et al., 2001). Pereira et al. (2010) found no seasonal variations in T. affinis density or a positive correlation with Sargassum spp. biomass; however, these authors observed a constant presence of juvenile T. affinis throughout the year in the phytal fauna of Sargassum, indicating that factors such as habitat availability may affect this gastropod’s abundance and recruitment. On the Piedade reef, climatic and hydrodynamic factors should be considered, because they may affect the dispersal of organisms in the reef pool, thereby causing a greater variability in the population distribution of T. affinis in the tide pool, in addition to the trampling of the reef flat.

The authors declare that there is no conflict of interests.

This study was partially funded by the National Council for Scientific and Technological Development [CNPq] Universal Announcement with process number 470424/2007-9.