Fig. 1 in Fig. 5 in Cymonomus curvirostris Sakai 1965

Applying Benthic Foraminiferal Assemblage to Evaluate the Coral Reef Condition in Dongsha Atoll lagoon Chienhsun Chen1,2,* and Hui-Ling Lin1,2 1Taiwan Ocean Research Institute, National Applied Research Laboratories, 80143 Kaohsiung, Taiwan 2Department of Oceanography, National Sun Yat-sen University, 80424 Kaohsiung, Taiwan (Received 29 March 2017; Accepted 27 June 2017; Published 20 July 2017; Communicated by Benny K.K. Chan) Chienhsun Chen and Hui-Ling Lin (2017) Coral reefs in the South China Sea are threatened by environmental changes and anthropogenic disruptions. Foraminifera, a group of unicellular shelled protists, has been considered a reliable indicator of water quality and reef health. However, this indicator has not yet been used to study coral reefs in the South China Sea. In this study of foraminifera in the lagoon of Dongsha Atoll, both the assemblage diversity and the functional group composition were investigated. The FORAM index (FI), a numerical indicator based on functional group composition, was used to evaluate the condition of the coral reef of the Atoll lagoon. A typical assumption is that FI = 4 is the minimum index value corresponding to a suitable environment for the growth of calcifying organisms. Environments with FI values between 2 and 4 are unsuitable to marginal for recovery of coral communities after a mortality event. Data were recorded regarding a total of 287 foraminiferal species. Approximately 68% of the tests belonged to the porcelaneous taxa; the Quinqueloculina, Triloculina, and Pseudomassilina species were well represented. Hyaline foraminifera were less abundant than porcelaneous foraminifera; the agglutinated foraminifera were the least abundant. Multivariate analyses revealed four clustering groups. The functional groups were classified according to FI’s definition; the heterotrophic group was the most abundant, accounting for 82% of foraminiferal abundance. The symbiont-bearing group was the second most abundant, with a relative abundance of 12%; and the stress-tolerant group was the least abundant. Two samples had FI values slightly higher than 4; Amphisorus and Calcarina were dominant, but the other symbiont-bearing foraminifera had lower values. These high FI values can be associated with seagrass meadows or relict shells in an altered environment. Eighty percent of the sediment samples had FI values lower than 4, indicating that most of the benthic habitats in the Dongsha lagoon are not suitable for coral growth and reef recovery. The present findings provide an alternative evaluation method based on foraminiferal assemblages for future studies on the environmental changes of coral reefs. Key words: Benthic foraminifera, FORAM index, Coral reef, Dongsha Atoll, South China Sea. *Correspondence: E-mail: [email protected] BACKGROUND Coral reefs are vital coastal ecosystems that provide crucial services and goods to maritime tropical and subtropical nations (Costanza et al. 1997). However, coral reefs are rapidly declining globally because of various stressors, such as climate change and anthropogenic disruptions (reviewed in Richmond 1993; Hoegh-Guldberg et al. 2007). Furthermore, 15% of the coral reefs worldwide are currently predicted to be lost in 2040 years unless more effective management is implemented (Bellwood et al. 2004; Knowlton and Jackson 2008). Foraminifera are the most abundant shelled microorganisms, which substantially contribute to the carbonate sediments in shallow-water environments. Owing to their prolific abundance, short reproductive cycle, and sensitivity to environmental changes, benthic foraminifera have been used as exceptional bioindicators of marine coastal environments. They have been used to evaluate Zoological Studies 56: 20 (2017) doi:10.6620/ZS.2017.56-20 1 a wide range of anthropogenic pollution (e.g. Resig 1960; Watkins 1961) and to assess the consequences of organic waste discharges (e.g. Alve 1995; Schafer 2000; Samir and El-Din 2001; Yanko et al. 2002). In the past decade, benthic foraminiferal assemblages have been employed for the rapid assessment of coral reef conditions (reviewed in Cooper et al. 2009). Hallock et al. (2003) developed a numerical index to associate water quality with reef health, known as the Foraminifera in Reef Assessment and Monitoring (FORAM) index (FI). The FI uses benthic foraminiferal assemblages to quantify the suitability of an environment to support the prolific growth of calcifying organisms that host endosymbiotic algae, such as zooxanthellate corals and larger benthic foraminifera. These endosymbionts aid in the calcification and food production for foraminifera just as zooxanthellae aid corals (Hallock 2000). However, endosymbionts constrain the growth of symbiont-bearing foraminifera and corals within the photic zone, particularly in relatively clear and oligotrophic waters (Hallock 1999). When the nutrient supply increases, fleshy algae and sponges dominate the benthic community. Consequently, foraminiferal assemblages (which consist of small and fast-growing species) can be measured to quantify environmental quality with respect to reef health (Reymond et al. 2012). Previous applications of the FI have revealed the advantages of foraminifera as bioindicators of reef conditions in the Aegean Sea (Koukousioura et al. 2011), Brazil (Barbosa et al. 2009; Barbosa et al. 2012), Colombia (Velásquez et al. 2011), Florida (Carnahan et al. 2009), Moorea, French Polynesia (Fajemila et al. 2016), and the Great Barrier Reef (Schueth and Frank 2008; Uthicke and Nobes 2008). However, the FI has not yet been applied to evaluate coral reefs in the South China Sea (SCS). Dongsha Atoll is a coral atoll in the northern SCS. The atoll’s circular-shaped lagoon is 16 km in width and provides a wide array of habitats, including seagrass meadows and coral reefs. It has been recognized for its biodiversity and ecological significance in the northern SCS (Morton 2002; Dai 2006). Early surveys of the region examined a limited number of sites close to the Dongsha Islands. For example, a survey conducted in 1994 recorded abundant coral and fish, high coral coverage (80%-95%), and welldeveloped communities of large Acropora and Porites colonies around Dongsha Island (Chen et al. 1995; Dai et al. 1995). However, owing to the 1998 worldwide bleaching event and frequent destructive fishing practices, coral reefs in the lagoon of Dongsha Atoll (hereafter referred to as the Dongsha lagoon) were degraded (Morton 2002; Dai 2006). Regular ecological surveys were conducted after the degradation, but the recovery of the reef in the Dongsha lagoon remains debatable. An effective assessment of the status of this coral reef is required. Therefore, this study investigated the assemblages of benthic foraminifera to provide an alternative assessment of the benthic environment in the Dongsha lagoon. The present work is the first investigation to study the benthic foraminifera in the Dongsha lagoon. This study aims (1) to assess the composition and assemblages of the benthic foraminifera in the Dongsha lagoon and (2) to use fast-responding foraminifera to assess the present reef-building status of the lagoon environment. MATERIALS AND METHODS Sampling Surface sediments from the Dongsha lagoon were collected in September 2009, in May and September 2010, and in August 2013. Evenly distributed sampling stations in the lagoon were originally planned; however, some stations were relocated because of the uncertain weather and sea conditions. Field sampling and laboratory preparation of the samples were performed following the methods described by Hallock et al. (2003) and Narayan and Pandolfi (2010), with minor modifications. Typically, scuba divers scooped the upper 1-2 centimeters of the surface sediments into a plastic bag. Each sample was treated with 4% neutral formaldehyde and stored in a plastic bag until it was transported to the laboratory. Figure 1 and table 1 present the sampling sites and the data obtained from the samples. Laboratory preparation In the laboratory, each sample was examined, and any algae and sea grass were removed. Each sample was freeze-dried and agitated thoroughly. According to the preliminary determinations of ideal quantitative counts, 200-300 individuals, approximately 10 g of freeze-dried sediment subsample was wet-sieved by using deionized water over a 125-μm mesh sieve. The samples page 2 of 16Zoological Studies 56: 20 (2017) Fig. 1. Map of Dongsha Atoll showing main geomorphological features and sampling sites in this study. ■ Dongsha Island; Back and fore reefs; □ Reef flat; ● Sampling sites. N Table 1. Information on benthic foraminifera in the Dongsha lagoon, including GPS and depth of the sediment sample, relative percentage of foraminiferal taxa with different shell textures, diversity index, relative percentage of foraminiferal taxa of different functional groups and the FORAM index. Shell texture: relative percentage of foraminifera with porcelaneous (P), hyaline (Y), and agglutinated (A) shell texture. Diversity index: Margalef richness index (d), Simpson dominance index (D), and Shannon-Wiener index (H). Functional group: symbiont-bearing (S), stress-tolerant (O), and other small and heterotrophic taxa (H) Sample GPS Depth Identified Shell texture (%) Diversity index Functional group (%) FORAM (latitude, longitude) (m) species P Y A d D H S O H index A 20°42.59'N, 116°45.91'E 5 69 76.8 21.9 1.3 12.4 0.06 3.6 35.9 5.1 59.1 4.8 B 20°43.38'N, 116°47.78'E 12 68 60.6 34.6 4.8 10.6 0.04 3.6 2.7 9.1 88.3 2.1 C 20°44.48'N, 116°50.68'E 13 77 71.9 17.3 10.8 10.8 0.05 3.4 13.6 4.3 82.1 3.0 D 20°41.65'N, 116°47.17'E 15 44 48.3 45.2 6.6 7.0 0.04 3.4 2.2 10.0 87.8 2.1 E 20°41.95'N, 116°50.53'E 15 98 77.9 22.1 0 17.2 0.02 4.2 11.7 4.3 84.0 2.9 F 20°42.48'N, 116°52.75'E 5 86 84.8 11.1 4.0 14.9 0.03 4.0 5.7 5.1 89.2 2.4 G 20°39.80'N, 116°46.21'E 10 88 40.0 60.0 0 15.4 0.06 3.7 44.3 9.6 46.1 5.4 H 20°40.52'N, 116°48.42'E 5 93 72.1 12.4 15.5 16.2 0.03 4.0 3.4 3.1 93.4 2.2 I20°40.65'N, 116°52.98'E 12 93 60.8 31.2 8.0 13.3 0.04 3.8 2.1 3.5 94.4 2.1 J 20°37.85'N, 116°50.47'E 12 101 82.5 17.1 0.3 17.4 0.02 4.2 3.8 6.3 89.8 2.2 Average 81.7 67.5 27.4 5.1 13.5 0.04 3.8 12.5 6.0 81.4 2.9 page 3 of 16Zoological Studies 56: 20 (2017) separated from the >125-μm fraction were dried in an oven at 50°C overnight and stored at room temperature. All benthic foraminiferal specimens from the >125-μm fraction were handpicked from each subsample and placed on microslides for identification under a standard dissecting microscope. We examined the total foraminiferal assemblages in each sediment subsample. The total assemblages provided information about the overall conditions that had accumulated over time. If the number of foraminiferal specimens was lower than the ideal quantitative count for a sample, additional sediment subsamples were weighed and processed as previously described. The taxonomic assignments of the benthic foraminifera were determined under a dissecting microscope. The supraspecific identification was based on their test characteristics, such as wall material and structure, as well as chamber arrangement and aperture position, following the classification criteria of Loeblich and Tappan (1988). Taxonomic identification was performed using taxonomic monographs of the SCS (Zheng and Zheng 1978; Zheng 1979; Zheng 1980), Australian (Nobes and Uthicke 2008; Debenay 2012) and Indo-Pacific regions (Cushman 1910, 1921; Jones 1994). Intact specimens that showed no signs of reworking and damage were included in the following analyses. A few abraded specimens (1-18 individuals per sample station) were found in this study. For each representative specimen, several images were captured from different focal lengths by using a digital camera (Canon EOS 7D) mounted on a stereoscopic microscope (Leica M205). The focus depths were subsequently enhanced and multiple images were merged digitally into one picture by using Helicon Focus (Helicon Soft Ltd.), as illustrated in figure 2. All samples and specimen illustrations were deposited in the micropaleontological collection of the Marine Core Repository and Laboratory at the Taiwan Ocean Research Institute, National Applied Research Laboratories. Analysis of foraminiferal assemblages and the biodiversity index The variability in the foraminiferal community was quantitatively evaluated in terms of the species relative abundance (RA); the number of identified species; and diversity indices, including species richness (Margalef richness index, d), dominance (Simpson dominance index, D), and diversity (Shannon-Wiener diversity index, H). These indices were calculated for each sediment sample. The raw count data were first tabulated and transformed to RA percentages by dividing the number of individuals of a species (n) by the total number of individuals (N), where RA = 100 × n/N. Each RA was used to conduct basic descriptive statistical analysis, including faunal composition and percentages of taxa with different wall compositions. Margalef richness index (d), which estimates species richness, and is independent of the sample size, was calculated as follows: d = (S - 1)/ln (N) where S is the number of species (Margalef 1958). Simpson dominance index is based on the probability that two individuals randomly selected from a sample belong to the same species (Simpson 1949). The Simpson dominance index, ranging from 0 (equal distribution of individuals among species) to 1 (dominance of one species in the community), was calculated as follows: D = Σ (Pi)2 where Pi is the proportion of individuals belonging to species i. The Shannon-Wiener diversity index (H) was calculated on the basis of the RA data: H = −Σ [Pi × ln(Pi)] It is a measure of heterogeneity considering the evenness of species abundance (Shannon 1948; Murray 1991). Differences in diversity indices between sediment samples were analyzed using pairwise permutation tests with 9999 random iterations. Calculations and statistical tests were conducted using PAST 3.15 (Hammer et al. 2001). Functional groups and the FORAM index The FI, initially described by Hallock et al. (2003) for assessing whether a benthic environment is hospitable to symbiont-bearing organisms (i.e., corals and reefal foraminifera), was calculated. The FI was calculated according to the foraminiferal composition by arranging the foraminiferal genera into three functional groups, namely: (1) taxa of large foraminifera that hosted algal symbionts, (2) pollution-tolerant opportunistic foraminifera that dominated high-stress environments, and (3) other small heterotrophic foraminifera. The FI was calculated as follows: page 4 of 16Zoological Studies 56: 20 (2017) Fig. 2. Foraminifera from the Dongsha lagoon. 1. Calcarina mayori Cushman; 2. Pseudohauerina orientalis (Cushman); 3. Pseudolachlanella eburnea (d’Orbigny); 4. Bolivina variabilis (Williamson); 5. Amphisorus hemprichii Ehrenberg, a: lateral view, b: apertural view; 6. Spiroloculina attenuate Cushman, lateral view; 7. Dendritina striata Hofker, a: lateral view, b: aperatural view; 8. Miliolinella subrotunda (Montagu), a: lateral view, b: apertural view; 9. Elphidium crispum (Linn), a: lateral view, b, apertural view; 10. Cymbaloporetta bradyi (Cushman), a: dorsal view, b: ventral view; 11. Rosalina globularis d’Orbigny, a: dorsal view, b: ventral view; 12. Pyrgo denticulata (Brady), a: lateral view, b: apertural view; 13. Pseudomassilina pacificiensis Cushman, a: lateral view, b: apertural view; 14. Triloculina wiesneri Le Calvez and Le Calvez, a: lateral view, b: aperatural view; 15. Spirillina grosseperforata Zheng, lateral view; 16. Fijiella simplex (Cushman), a: apertural view, b: lateral view; 17. Reussella pacifica Cushman and McCulloch, a: apertural view, b: lateral view; 18. Textularia agglutinans d’Orbigny, side view. Scale bar = 100 µm. 13 4 5a 5b 7a 7b 8a 8b 9a 9b 10a 10b 11a 11b 12a 12b 13a 13b 14a 14b 15 16a 16b 17a 17b 18 6 2 page 5 of 16Zoological Studies 56: 20 (2017) FI = (10 × Ps) + (Po) + (2 × Ph) where Ps, Po, and Ph represent the proportion of symbiont-bearing, stress-tolerant, and other heterotrophic taxa, respectively. The FI is a singlemetric indicator of water quality that has already been applied in tropical coral reef communities and subtropical estuarine environments (reviewed in Hallock 2012; Reymond et al. 2012). A coral reef with FI > 4 is an oligotrophic environment supporting calcifying mixotrophs. However, an FI between 2 and 4 indicates a marginal environment for calcifying foraminifera, and FI < 2 indicates environmental conditions that support substantial populations of stress-tolerant foraminifera (Hallock et al. 2003). Multivariate statistics For determining the structures of the foraminiferal assemblages, RA of a subset of the most abundant 23 foraminiferal genera that accounted for approximately 85% of the total population of the foraminifera was used. Q-mode hierarchical clustering techniques were applied to retrieve information and discover structural entities within complex data sets (Parker and Arnold 1999). A nonmetric multidimensional scaling (nmMDS) analysis was carried out to visualize and refine the grouping pattern among samples. The betweengroup differences in foraminiferal composition were tested by analysis of similarity (ANOSIM). The taxa contributing most to the dissimilarity among groups were identified by similarity percentage analysis (SIMPER). Distance and similarity measurements of the multivariate analyses were calculated based on the Bray-Curtis distance. For cluster grouping, a paired grouping algorithm with a bootstrapping procedure of 500 iterations was executed to evaluate the stability of the clustering results. The ANOSIM was tested with 9999 permutations. Calculations and statistical tests were conducted using PAST 3.15 (Hammer et al. 2001). RESULTS Benthic foraminiferal assemblage A total of 4,935 benthic foraminifera were recognized, which accounted for 287 species belonging to 111 genera distributed among the orders Astrorehizida, Lituolida, Textulariina, Miliolida, Spirilinida, Lagenida, Robertinida, Buliminida, and Rotaliida (Appendix 1). Of the 287 identified species, only 63 (22%) species consistently occurred and were found in more than half of the sediment samples (Appendix 1). When all foraminiferal specimens of the 10 samples in the Dongsha lagoon were combined, approximately 98% of the foraminiferal species had RA values of less than 3%. Only five species, Quinqueloculina bosciana (RA = 3.5%), Pseudohauerina orientalis (RA = 3.4%), Fijiella simplex (RA = 3.2%), Q. debenayi (RA = 3.2%), and Textularia agglutinans (RA = 3.1%), had RA values exceeding 3%. Porcelaneous foraminifera were predominant throughout the Dongsha lagoon (Appendix 1). For example, among the consistent species, 76.2% were porcelaneous foraminifera (48 species), belonging to Pseudohauerina of Brebinidae; Planispirinella and Vertebralina of Fischerinidae; Miliolinella, Pseudolachlanella, Pseudomassilina, Pyrgo, Quinqueloculina, Schlumbergerina, Sigmamiliolonella, and Triloculina of Hauerinidae; Dendritina of Peneroplidae; Amphisorus, Parasorites, and Sorites of Soriitae; and Spiroloculina of Spiroloculinidae. Fourteen hyaline species were consistent, including those belonging to Ammonia and Neorotalia of Rotaliidae, Fijiella of Reussellidae, Epistomaroides of Alfredinidae, Cibicides of Cibicididae, Cymbaloporetta of Cymbaloporidae, Elphidium of Elphidiidae, Heterostegina of Nummulitidae, and Rosalina of Rosalinidae. Only one agglutinated species, belonging to the genus Textularia, consistently occurred. Porcelaneous foraminifera were the most abundant foraminifera in the Dongsha lagoon, with RAs of approximately 40%-84.8% in our sediment samples (Table 1). Quinqueloculina was the most abundant genus (RA = 28.1%) and was well-represented with Q. bosciana as the most abundant species (RA = 3.5%), followed by Q. debenayi (RA = 3.2%) and Q. parkeri (RA = 2.9%). Triloculina and Pseudomassilina were the second and third most abundant porcelaneous genera with RAs of 6.4% and 5.2%, respectively. Hyaline foraminifera, with an average RA of 27.4%, were less abundant than porcelaneous foraminifera. Elphidium and Fijiella were the most abundant hyaline taxa, with RAs of 3.4% and 3.2% respectively. Agglutinated taxa were the least abundant (average RA = 5.1%, Table 1) and were recorded in only a total of 21 species. Textularia was the most abundant agglutinated genus (RA = 3.9%) and was well-represented with T. agglutinans (RA = 3.1%). page 6 of 16Zoological Studies 56: 20 (2017) Biodiversity index In each sample, 44-101 species were identified, with a mean of 81.7 (Table 1). The least species, occurring in sample D, was the smallest value from the rest, but it was not a significant outlier (p > 0.05; Grubbs test, Z = 2.29). The Margalef richness index of the foraminiferal assemblages ranged between 7.0 and 17.4 (Table 1). The Margalef richness index of the foraminiferal assemblages was the highest in samples J and E and was on an average approximately 1.2 times higher than that in samples A, B, C, and D. Pairwise permutation testing revealed that the sediments from the east and southeast regions of the Dongsha lagoon had high species richness (p < 0.001 with Bonferroni correction in all pairwise permutation tests, Fig. 3a). The Simpson dominance index ranged between 0.02 and 0.06 (Table 1). The north and west regions of the Dongsha lagoon showed high dominance index values of benthic foraminifera. Species, such as Amphisorus hemprichii (RA = 20.3% in sample A), Calcarina spp (RA = 29.7% in sample G), and Quinqueloculina debenayi (RA = 12.5% in sample C) were dominant at these stations. However, the RA of each foraminiferal species was lower than 10% in the remaining samples. The Shannon-Wiener diversity index ranged between 3.4 and 4.2 (Table 1). Pairwise permutation analyses revealed higher foraminiferal diversity in samples collected from the east and southeast regions, such as samples J and E; the diversity was approximately 1.2 times higher than that in samples collected from the north and west regions (samples A, B, C, and D; Fig. 3c). Functional groups and the FORAM index According to the functional groups defined by Hollock et al. (2003), the heterotrophic group was predominant, which comprised approximately 81.4% of the foraminifera in the benthic environment of the Dongsha lagoon (Table 1). Among the 93 heterotrophic genera, 10, including Fijiella, Miliolinella, Pseudohauerina, Pseudomassilina, Quinqueloculina, Reussella, Spirillina, Spiroloculina, Textularia, and Triloculina, had RA values > 5% at more than two stations. Symbiont-bearing foraminifera were not abundant, with an average RA of 12.5% (Table 1). Thirteen genera were classified as symbiontbearing foraminifera, including Alveolinella of Fig. 3. Diversity index of benthic foraminifera from surface sediments of the Dongsha Atoll lagoon: (a) Margalef richness index; (b) Simpson dominance index; and (c) Shannon - Wiener diversity index. Error bars indicate the upper and lower limits of 9999 bootstrap iterations. Horizontal lines under the x-axis indicate no significant differences among the sampling sites (p > 0.001, pairwise comparison with 9999 random permutation iterations). Sampling sites refer to table 1 and figure 1. (a) (b) (c) page 7 of 16Zoological Studies 56: 20 (2017) Alveolinidae; Amphistegina of Amphisteginidae; Baculogypsina and Calcarina of Calcarinidae; Heterostegina and Operculina of Nummulitidae; Coscinopira and Dendritina of Peneroplidae; Neorotalia of Rotaliidae; and Amphisorus, Marginopora, Parasorites, and Sorites of Soritidae (Table 2). Only Calcarina had RA values >5% in two samples. Stress-tolerant foraminifera constituted the least abundant group, with an average RA of 6% (Table 1); this group included Bolivina and Bolivinellina of Bolivinidae, Elphidium of Elphidiidae, Ammonia of Rotaliidae, and Ammobaculite of Lituolidae (Table 2). The FI ranged between 2.1 and 5.4 and was typically low in the Dongsha lagoon (Table 1). Eighty percent of the samples had FI values < 4. The FI exceeded 4 for samples A (FI = 4.8) and G (FI = 5.4), where Amphisorus (RA = 24.1%) and Calcarina (RA = 29.6%) were dominant. Multivariate analyses Cluster analysis comparing the compositions and RAs of foraminiferal species in all samples revealed four clusters at 65% similarity, with a high cophenetic coefficient (0.90) and robust support (bootstrap value > 50%; Fig. 4). However, the results of cluster analysis were not consistent with the geological affinity of the samples. Clusters 1 and 2 were observed in samples G and A, respectively. These samples were characterized by median to coarse sand and high levels of symbiont-bearing foraminifera, such as A. hemprichii and Calcarina species. The prevalence of other symbiont-bearing foraminifera, including A. sauronensis, Heterostegina depressa, Parasorites orbitolitoides, and Sorites orbiculus, was minimal. Quinqueloculinids constituted approximately 13.6% and 24.9% of clusters 1 and 2, respectively. Other thick-shelled miliolids, such as Miliolinella subrotunda, Pseudohauerina involute, Pyrgo depressa, Spiroloculina antillarum, and Triloculina wiesneri, occurred in minimal proportions. Opportunistic taxa, such as Elphidium, were abundant in sample G (RA = 7.1%). Cluster 3 mostly consisted of samples from the northeast region of the Dongsha lagoon, including samples C, E, F, H, and J. Sixty-five percent of similarity and a moderate bootstrap support (56%) indicated potential subgroups within this cluster. This cluster included a diverse assemblage of 222 species of benthic foraminifera. It was characterized by a high abundance of Quinqueloculina (RA = 31.7%-45.8%), which was represented by 55 species, and it constituted 41.4% of all specimens collected from the sampling stations. The agglutinated taxon of Textularia, accounting for 6.8% of all specimens, was the second most predominant taxon. The proportion of symbiont-bearing foraminifera, such as Coscinospira hemprichii and Dendritina striata, was moderate in sample C (RA = approximately 6%) but minimal in the other samples (RA = 0%- 2.2%). The prevalence levels of other miliolids, Table 2. Generic categorization of functional groups of foraminifera in the Dongsha lagoon. Stars indicate the selected 23 genera of benthic foraminifera included in cluster analyses Functional groups Genera of benthic foraminifera Symbiont-bearing Alveolinella, Amphisorus*, Amphistegina, Baculogypsina, Calcarina*, Coscinospira*, Dendritina*, Heterostegina, Marginopora, Neorotalia, Operculina, Parasorites, Sorites Opportunistic Ammobaculites*, Ammonia, Bolivina*, Bolivinellina, Elphidium* Heterotrophic Acervulina, Acupeina, Adelosina, Alliatina, Alliatinella, Ammomassilina, Anomalinella, Anomalinoides, Articulina, Ashbrookia, Baggina, Bronnimannia, Buliminoides, Cancris, Caribeanella, Cibicides, Cibicidoides, Cibicorbis, Clavulina, Conicospirillinoides, Cornuspira, Cymbaloporella*, Cymbaloporetta, Discorbinella, Epistomaroides, Eponides, Fijiella*, Fissuripolymorphina, Fontbotia, Fursenkoina, Gyroidina, Hanzawaia, Hauerina, Haynesina*, Heterolepa, Hyperammina, Inaequalina, Laryngosigma, Lenticulina, Melonis, Miliammina*, Miliola, Miliolinella*, Millettiana, Mychostomina, Neocassidulina, Neoconorbina, Neoeponides, Nodobaculariella, Nonion, Nonionoides, Nummulopyrgo, Pararotalia, Parrina, Planispirillina, Planispirinella, Planogypsina, Planorbulinella, Planulina, Poroeponides, Porosononion, Pseudohauerina*, Pseudolachlanella*, Pseudomassilina*, Pseudotriloculina, Pyrgo*, Pyrgoella, Quinqueloculina*, Reophax, Reussella*, Rosalina*, Rotorbis, Saintclairoides, Schlumbergerina, Sigmamiliolinella, Sigmavirgulina, Sigmoilinella, Sigmoilopsis, Siphogenerina, Siphonina, Siphotextularia, Spirillina*, Spiroloculina*, Spirosigmoilina, Textularia*, Triloculina*, Triloculinella, Ungulatelloides, Uvigerina, Valvulineria, Vertebralina, Wiesnerella, Zoyaella page 8 of 16Zoological Studies 56: 20 (2017) such as Pseudohauerina orientalis (RA = 1.4%- 5.3%) and Pseudomassilina robusta (RA = 0%- 11.4%) and species of Miliolinella (RA = 0%- 2.8%), Spiroloculina (RA = 0%-3.1%), and Triloculina (RA = 0%-2.8%) were moderate to minimal. Opportunistic taxa, such as Bolivina and Elphidium, occurred in minimal quantities. Cluster 4 comprised samples B, D, and I, and consisted of 153 species of benthic foraminifera. The thick-shelled miliolids were prominent in this cluster. The prevalence of quinqueloculinids was 13.7% at these sites, followed by Miliolinella (7.1%), Reussella (RA = 6.9%), Pseudomassilina (RA = 6.8%), Triloculina (RA = 6.2%), Spirillina (RA = 6.0%), Fijiella (RA = 5.9%), Spiroloculina (RA = 5.8%), and Pseudohauerina (RA = 5.1%). Symbiont-bearing taxa were rare and accounted for only 2% of all foraminifera. An nmMDS ordination plot based on same data set grouped the sampling stations in approximately the same manner as that of cluster analysis (Fig. 5). On an nmMDS plot the distance between two points corresponds to their similarity in composition. Different clustering groups were distributed in various quadrants of the plot with a fair goodness of fit (Fig. 5; stress = 0.10). Results of the ANOSIM analyses showed significant differences in foraminiferal assemblage composition between clusters 3 and 4 (R = 0.98, p = 0.02 with 9999 permutations). Because each of clusters 1 and 2 had only one sediment sample, clusters 1 and 2 were excluded from the ANOSIM test. Results of SIMPER analyses indicated that overall dissimilarities ranged from 49% to 59.9% (Table 3). In most of the pairwise comparisons, abundant genera were the taxa contributing most to the between-group differences. For example, Calcarina and Amphisorus contributed most to the between-group dissimilarity of clusters 1 and 2, respectively (Table 3). Fig. 4. Dendrogram of the sampling sites plotted by cluster analysis by using Bray-Curtis similarity based on the subset dataset of 23 foraminiferal genera. Values on the dendrograms indicate the > 50% supporting value of 500 bootstrap iterations. Amphisorus Calcarina Coscinospira Dendritina Miliammina Textularia Pseudohauerina Miliolinella Pseudolachlanella Pseudomassilina Pyrgo Quinqueloculina Triloculina Spiroloculina Spirillina Fijiella Reussella Cymbaloporetta Haynesina Rosalina Ammobaculites Bolivina Elphidium Similarity 0.42 0.48 0.54 0.60 0.66 0.72 0.78 0.84 0.90 0.96 G C E F J H A B D I Cluster 1 Relative abundance : 0 - 0.9 : 1 - 4.9 : 5 - 9.9 : 10 - 19.9 : 20 - 39.9 : >40 Cluster 3 Cluster 2 Cluster 4 56 77 100 page 9 of 16Zoological Studies 56: 20 (2017) Appendix 1. Foraminiferal species list. Stars indicate consistently occurring species that were found in more than half of sampling locations. (download) page 16 of 16Zoological Studies 56: 20 (2017)