INTRODUCTION
Microplankton organisms are the most important component of the plankton and play a significant role in the functioning of the microbial food loop, especially in terms of energy flow and element cycling in many aquatic ecosystems (Montagnes et al., Reference Montagnes, Berger and Taylor1996; Dolan & Simek, Reference Dolan and Simek1997; Jiang et al., Reference Jiang, Xu, Hu, Zhu, Al-Rasheid and Warren2011a, Reference Jiang, Xu, Al-Rasheid, Warren, Hu and Songb, Reference Jiang, Xu, Zhang, Zhu and Al-Rasheid2012a, Reference Jiang, Zhang, Zhu, Al-Rasheid and Xub; Xu et al., Reference Xu, Jiang, Al-Rasheid, Song and Warren2011a, Reference Xu, Jiang, Al-Rasheid, Al-Farraj and Songb). The microalgae are responsible for the primary production in most aquatic habitats; microzooplanktons transfer these productions to higher trophic levels in the food chain (Tillmann, Reference Tillmann1998, Reference Tillmann2004; Gomez & Gorsky, Reference Gomez and Gorsky2003). Furthermore, some bloom-forming species are harmful to the other microorganisms and often result in red-tide events in the marine ecosystems (Tillmann, Reference Tillmann2004; Xu et al., Reference Xu, Song, Warren, Al-Rasheid, Al-Farraj, Gong and Hu2008, Reference Xu, Min, Choi, Zhu, Jiang and Al-Rasheid2010).
The planktonic ciliated protozoa as an important component of microzooplankton, with short generation time, can react rapidly to short-term variation in food conditions in case of rapid phytoplankton growth (Admiraal & Venekamp, Reference Admiraal and Venekamp1986; Montagnes et al., Reference Montagnes, Berger and Taylor1996; Jeong et al., Reference Jeong, Shim, Lee, Kim and Koh1999). Many investigations have reported high numbers of ciliate species during bloom events, suggesting that planktonic ciliates play crucial roles in suppressing or shortening blooming events of microalgae (Admiraal & Venekamp, Reference Admiraal and Venekamp1986; Bochstahler & Coats, Reference Bochstahler and Coats1993; Agatha & Riedel-Lorjé, Reference Agatha and Riedel-Lorjé1998; Montagnes & Lessard, Reference Montagnes and Lessard1999; Tillmann, Reference Tillmann2004). This assumption is supported by laboratory studies showing the capability of different ciliate species to feed and grow on bloom-forming algal species (Verity, Reference Verity1985; Bernard & Rassoulzadegan, Reference Bernard and Rassoulzadegan1990; Stoecker & McDowell Cappuzzo, Reference Stoecker and McDowell Capuzzo1990; Stoecker & Michaels, Reference Stoecker and Michaels1991; Montagnes et al., Reference Montagnes, Berger and Taylor1996; Dolan & Simek, Reference Dolan and Simek1997; Strom & Morello, Reference Strom and Morello1998; Jeong et al., Reference Jeong, Shim, Lee, Kim and Koh1999; Kamiyama & Arima, Reference Kamiyama and Arima2001; Tang et al., Reference Tang, Jakobsen and Visser2001; Pedersen & Hansen, Reference Pedersen and Hansen2003). As regards the interspecies interactions between the microplanktonic grazers and the microalgae, however, further investigations on temporal distribution patterns of microplankton communities using multivariate-statistical approaches are still needed although a few relevant researches have been reported (Tillmann, Reference Tillmann2004).
In the present study, the temporal pattern of microplankton communities and interspecies correlations between ciliates and microalgae were analysed, using a range of multivariate statistical methods, based on a dataset of microplankton communities, which was collected biweekly at five sampling sites in Jiaozhou Bay near Qingdao, northern China, during a 1-year cycle (June 2007–May 2008). Our study asks the following questions: (1) how do the distribution patterns of microplankton communities change in an annual cycle?; (2) what are their relationships with environmental changes?; and (3) what are the interactions between planktonic ciliates and microalgae in marine ecosystems?
MATERIALS AND METHODS
Study sites
Jiaozhou Bay is a semi-enclosed basin near Qingdao, northern China. It covers an area of about 390 km2 with an average depth of 7 m and is connected to the Yellow Sea via a narrow opening about 2.5 km wide. Five sampling sites (A–E) were selected in this Bay (Figure 1).
Sampling stations of microplankton in Jiaozhou Bay.
Sampling, fixation, measurements, identification and enumeration
The study was conducted during June 2007 to May 2008 in Jiaozhou Bay, northern China (Jiang et al., Reference Jiang, Xu, Hu, Zhu, Al-Rasheid and Warren2011a, Reference Jiang, Xu, Al-Rasheid, Warren, Hu and Songb). The sampling strategy followed that described by Jiang et al. (Reference Jiang, Xu, Hu, Zhu, Al-Rasheid and Warren2011a, Reference Jiang, Xu, Al-Rasheid, Warren, Hu and Songb).
Salinity (Sal), pH, and dissolved oxygen concentration (DO) were measured in situ, using a multi-parameter sensor (MS5, HACH). Samples for nutrient analyses were preserved immediately upon collection by placing them at –20°C in the dark. Concentrations of soluble reactive phosphorus (SRP), ammonium nitrogen (NH3-N), nitrate nitrogen (NO3-N) and nitrite nitrogen (NO2-N) were determined using a UV-visible spectrophotometer (DR-5000, HACH) according to the Standard Methods for the Examination of Water and Wastewater (APHA, 1992). For enumeration of ciliates, diatoms and dinoflagellates, a 0.1 ml aliquot of each concentrated sample was placed in a Perspex chamber and counted under a light microscope at ×400-magnification. A total of 0.5 ml concentrated samples were counted and yielded a standard error of <8% of the mean values of counts. The protargol staining method for ciliates was performed according to the protocol of Montagnes & Humphrey (Reference Montagnes and Humphrey1998). Scanning electron microscopy was used to identify the microalgal species hard to distinguish by light microscope. The cells were identified to the lowest taxonomic level possible based on the published references to keys and guides such as Hasle & Syvertsen (Reference Hasle, Syvertsen and Tomas1997), Steidinger & Tangen (Reference Steidinger, Tangen and Tomas1997) and Song et al. (Reference Song, Zhao, Xu, Hu and Gong2003).
Data analyses
Multivariate analyses were carried out using the PRIMER v6.1 statistical package (Clarke & Gorley, Reference Clarke and Gorley2006), the PERMANOVA+ for PRIMER (Anderson et al., Reference Anderson, Gorley and Clarke2008) and the statistical program SPSS (version 16.0). Bray–Curtis similarity matrices were computed on species-abundance data while the temporal patterns of communities were summarized using the submodule CAP (canonical analysis of principal coordinates) of PERMANOVA+ on Bray–Curtis similarities. Differences between groups of samples were tested by the submodule ANOSIM (analysis of similarity) (Clarke & Gorley, Reference Clarke and Gorley2006). The significance of ciliate–microalgae correlations was tested using the routine RELATE (Clarke & Gorley, Reference Clarke and Gorley2006). The routine BEST was used to determine the typical species for both the ciliate and the microalgal assemblages (Clarke & Warwick, Reference Clarke and Warwick1994). The multidimensional scaling (MDS) ordination was used to summarize species distribution on Bray–Curtis similarity (Clarke & Gorley, Reference Clarke and Gorley2006). RELATE/BIOENV analyses were used to reveal the correlations between temporal patterns of microplankton communities and environmental conditions (Clarke & Gorley, Reference Clarke and Gorley2006).
The best possible regression models were explored using the stepwise selection mode and the optimal model was estimated based on the statistical significance (high R 2, P < 0.05), using the SPSS software. Biotic data were fourth root-transformed, while abiotic data were log-transformed before analyses.
RESULTS
Taxonomic composition
The taxonomic composition, average abundance and occurrence of microplankton (planktonic ciliates and microalgae) assemblages observed during the study period are summarized in Table 1. A total of 64 ciliate species and 100 microalgae species (basically dinoflagellates and diatoms) were identified from 120 samples during the 1-year survey in Jiaozhou Bay, northern China. The BEST analysis showed that the microplankton communities were characterized by 14 ciliates and 18 microalgae respectively (Table 1).
List of the species of planktonic ciliates (Cili, ciliates) and microalgae (Dino, dinoflagellates; Dia, diatoms) from Jiaozhou Bay recorded in 120 samples including taxon type, annual average abundance and occurrence.
*, typical species determined by routine BEST within 120 microplankton samples; 1, + = 10; ++ = 10–100; ++ + = 100–1000; ++ + + = 1000–10000; ++ + ++ = over 10000.
Temporal variations of community structures
The temporal patterns of planktonic ciliate and microalgae assemblages in 1-year cycle were discriminated by using the submodule CAP (Figure 2). The first canonical axis separated the ciliate assemblages sampled in summer (on the right) from those in autumn and winter (on the left), while the second canonical axis discriminated the samples in spring (lower) from summer and winter (upper) (Figure 2A). The two canonical axes clearly separated the microalgae assemblages sampled in four seasons (Figure 2B). The ANOSIM test demonstrated significant differences between each pair of temporal groups in ciliates (R = 0.305, P = 0.001) and microalgae (R = 0.158, P = 0.001).
Canonical analysis of principal coordinates (CAP) on Bray–Curtis similarities from species-abundance data of two assemblages (ciliates and microalgae) in 120 samples from five sampling sites in Jiaozhou Bay during the annual cycle from June 2007 to May 2008.
Temporal ordinations of species distribution
The temporal patterns of species distribution within ciliated and microalgal assemblages are summarized in Figures 3, 4, 5 and 6, using the MDS ordination on Bray–Curtis similarity from the log-transformed species-abundance data.
Ordination of the typical species on the multi-dimensional scaling-diagram (B) and temporal variation in relative abundance (A) in spring. Sp I–III, group Sp I–III.
Ordination of the typical species on the multi-dimensional scaling-diagram (B) and temporal variation in relative abundance (A) in summer. Su I–V, group Su I–V.
Ordination of the typical species on the multi-dimensional scaling-diagram (B) and temporal variation in relative abundance (A) in autumn. Au I–VI, group Au I–VI.
In spring, the species distribution represented only three groups (group Sp I–III) (Figure 3). Group Sp I was the primary contributor in communities and consisted of 10 ciliated species (e.g. Strombidium globosaneum and Strombidium sulcatum) and 3 microalgae (e.g. Alexandrium tamarense and Navicula sp.) while group Sp II included two microalgal species (Ditylum brightwellii and Guinardia delicatula) with two ciliates (Strombidinopsis cheshiri and Leprotintinnus bottnicus) and group Sp III involved four microalgal species (e.g. Skeletonema costatum and Coscinodiscus subtilis) (Figure 3B). In this season, group Sp I dominated almost all samples except in early March during which group Sp II became the primary contributor with microalgae blooming but quickly suppressed in late March by ciliates. Finally, in April, group Sp I dominated the community with the cooperation of group Sp III (Figure 3).
In summer, the species distribution comprised five groups (group Su I–V) of which groups Su I and Su II with most dominant species were the primary contributors to the communities in terms of abundance compared to group Su III–V with low abundance (Figure 4). Group Su I mainly comprised some ciliates coming from group Sp I in spring (e.g. Pseudotontonia cornuta, Omegastrombidium foissneri and Strombidium conpressum) associated with diatom Navicula sp. (also coming from group Sp I) and dinoflagellate Prorocentrum micans (from Sp III) while group Su II comprised mainly microalgal species which were bursting out, e.g. Ceratium tripos, Dictyocha fibula and Coscinodiscus asteromphalus, or coming from species in spring such as Alexandrium tamarense (group Sp I) and Skeletonema costatum (group Sp III) (Figure 4B). It should be noticed that the communities in summer presented a clear continuity to the pattern in spring (Figures 3 & 4): in early May, groups Su I and IV, with the species mainly from group Sp I in spring, dominated the community in cooperation with group Su II (Figures 3 & 4). After that, group Su I and II exhibited such an interspecies relationship that they alternately dominated the communities during this season until group Su II almost occupied the community in the last sample (Figure 4B).
In autumn, the species distribution consisted of six groups (group Au I–VI). Group Au I, included the 19 most dominant species (e.g. ciliates Leprotintinnus bottnicus, Tintinnopsis bubulosoides and Strombidium styliferum; diatom Skeletonema costatum and dinoflagellate Prorocentrum lima) mainly coming from group Su I and II in summer, and was the primary contributor to communities. While group Au II comprised four species of which, diatom Navicula sp. and ciliate Strombidium sulcatum were coming from summer group Su I and Prorocentrum minimum from group Su IV. As regards group Au III–VI, this was composed of several less dominant species from group Su II and III (Figures 4B &5B). Although with community continuity in summer, however, the pattern of communities was different (Figures 4 & 5): group Au I dominated the most samples afterwards being replaced by group Au II, followed by group Au IV and VI in late autumn (Figure 5A).
In winter, the species distribution represented six groups (group Wi I–VI) (Figure 6B). Group Wi I comprising 13 dominant species, of which mainly ciliates from autumn group Au I (e.g. ciliates Strombidium acutum, Strombidium styliferum and Leprotintinnus bottnicus) associated with diatom Guinardia delicatula of group Au III, was the main contributor to communities in all samples (Figures 5 & 6) while group Wi IV included two dinoflagellate Prorocentrum minimum (from group Au II) and Prorocentrum lima (from group Au I) in cooperation with two ciliates (from group Au I) (Figures 5B & 6B). Furthermore, group Wi V comprised ciliate Stenosemella nivalis (group Au I) and diatom Coscinodiscus oculus-iridis (group Au VI) (Figures 5B & 6B); group II–VI comprised several species from autumn groups but in low abundance (Figures 5 & 6). Although the community in late November is very consistent with that of early November, the temporal pattern had significant differences with that in autumn (Figures 5A & 6A): group Wi I was always the main component in communities in cooperation with group Wi III, IV and V and predominated in late winter (Figures 5A & 6A). After that, the species distribution in spring represented clear species relationships with that in winter (Figures 3B & 6B): primary contributor group Sp I were merged by group Wi I–IV, e.g. ciliates Strombidium globosaneum (from group Wi I), Strombidium sulcatum (group Wi II) and Strombidium montagnesi (group Wi II) were associated with microalgae Alexandrium tamarense (group Wi I), Navicula sp. (group Wi I) and Prorocentrum minimum (group Wi IV). Furthermore, some species of group Wi I forming the group Sp II and III, for instance, microalgae Ditylum brightwellii and Guinardia delicatula associated with two ciliates in group Sp II; Coscinodiscus subtilis associated with the other three microalgae in group Sp III.
Ordination of the typical species on the multi-dimensional scaling-diagram (B) and temporal variation in relative abundance (A) in winter. Wi I–VI, group Wi I–VI.
The temporal succession process of 32 microplankton species (14 ciliates and 18 microalgae) is indicated in Figure 7 and highly consistent with the MDS ordination results in Figures 3, 4, 5 and 6. In this 1-year cycle, planktonic ciliates and microalgae formed a loop together. During each season or period, the one-way arrows show succession of species and the two-way arrows show their occurrence at the same time (Figure 7).
1-year loop of 14 dominant ciliate species and 18 microalgae species in abundance. Species abbreviations: A. senarius, Actinoptychus senarius; A. tamarense, Alexandrium tamarense; B. paxillifera, Bacillaria paxillifera; C. furca, Ceratium furca; C. tripos, Ceratium tripos; C. asteromphalus, Coscinodiscus asteromphalus; C. oculus-iridis, Coscinodiscus oculus-iridis; C. subtilis, Coscinodiscus subtilis; D. fibula, Dictyocha fibula; D. brightwellii, Ditylum brightwellii; D. sol, Ditylum sol; G. delicatula, Guinardia delicatula; L. bottnicus, Leprotintinnus bottnicus; N. sp., Navicula sp.; O. foissneri, Omegastrombidium foissneri; P. faurei, Parastrombidium faurei; P. lima, Prorocentrum lima; P. micans, Prorocentrum micans; P. minimum, Prorocentrum minimum; P. cornuta, Pseudotontonia cornuta; R. setigera, Rhizosolenia setigera; S. cosatum, Skeletonema costatum; S. nivalis, Stenosemella nivalis; S. acutum, Strombidium acutum; S. cheshiri, Strombidinopsis cheshiri; S. compressum, Strombidium compressum; S. globosaneum, Strombidium globosaneum; S. montagnesi, Strombidium montagnesi; S. styliferum, Strombidium styliferum; S. sulcatum, Strombidium sulcatum; T. antarctica, Tontonia antarctica; T. tubulosoides, Tintinnopsis tubulosoides.
Interspecies correlations between planktonic ciliates and microalgae
The Mantel test, by using RELATE analysis, revealed that there was a significant correlation between temporal variations in ciliated and microalgal assemblage structures (R = 0.218; P = 0.001).
In Table 2, the correlations between the abundances of 14 ciliates and 18 microalgae were obtained by linear regression and indicated that species-specific correlations existed between 11 ciliated species and 13 microalgal species. Among these, eight species (Strombidium globosaneum, Leprotintinnus bottnicus, Strombidinopsis cheshiri, Strombidium compressum, Stenosemella nivalis, Tontonia antarctica, Tintinnopsis tubulosoides and Strombidium sulcatum) were found correlated with two or more microalgal species in abundance. For example, S. globosaneum, was not only positively correlated with two dinoflagellates (Alexandrium tamarense and Prorocentrum minimum) and two diatoms (Coscinodiscus asteromphalus and Guinardia delicatula) but also negatively correlated with the diatom Actinoptychus senarius. Only three species, however, Pseudotontonia cornuta, Strombidium montagnesi and Strombidium acutum, were associated with only one microalgal species (Table 2).
Linear regression analysis; abundance of protozoan grazer related to algival protists. F of the model is only shown when the model is multiple.
Results obtained by linear regression also indicated that the abundances of 11 microalgal species were correlated with that of 11 planktonic ciliates (Table 3). Six species (Ceratium tripos, Skeletonema costatum, Actinoptychus senarius, Guinardia delicatula, Bacillaria paxillifera and Coscinodiscus oculus-iridis) were found correlated with not only one species; however, five species (Prorocentrum lima, Alexandrium tamarense, Ditylum brightwellii, Dictyocha fibula and Ceratium furca) were associated with only one ciliate species (Table 3).
Linear regression analysis; abundance of algival protists related to protozoan grazer. F of the model is only shown when the model is multiple.
As specified above, the linear regression analyses on modelling the species-specific correlations between 12 planktonic ciliate species and 15 microalgal species are summarized as a microbial correlation web in Figure 8.
Microbial correlation web formed by 12 ciliated protozoa species and 15 microalgal species. See Figure 7 for all abbreviations.
Relationship between microplankton and environmental variables
The Mantel test demonstrated that the temporal variations in microplankton community structures were significantly correlated with those of environmental variables (P < 0.001).
For the temporal cycle, the correlations between microplankton abundances and environmental variables were established by multivariate biota–environment (BIOENV) analysis (Table 4). The results showed that the best matching with the planktonic ciliates occurred with the combination of temperature, salinity and NO3-N, while the best matching with microalgae occurred with the combination of temperature, pH and NO3-N. It was also notable that temperature and NO3-N were included in most correlations (Table 4).
Summary of results from biota–environment (BIOENV) analysis showing the 10 best matches of environmental variables with temporal variations in ciliate and microalgae abundances in Jiaozhou Bay from June 2007 to May 2008.
Tem, temperature; Sal, salinity; DO, dissolved oxygen; NO3-N, nitrate nitrogen; NO2-N, nitrite nitrogen.
DISCUSSION
So far, there has been a growing interest on the interspecies correlations of planktonic ciliated protozoan with microalgae especially in field studies, although many researches have analysed various aspects of growth and feeding of ciliate species in culture with many ingeniously designed to study the interactions of specifically ciliates with demonstrated microalgal species (Bernard & Rassoulzadegan, Reference Bernard and Rassoulzadegan1990; Dolan & Simek, Reference Dolan and Simek1997; Montagnes & Lessard, Reference Montagnes and Lessard1999; Granéli & Johansson, Reference Granéli and Johansson2003; Clough & Strom, Reference Clough and Strom2005). For example, several tintinnids and non-loricate ciliates (Heinbokel, Reference Heinbokel1978; Jeong et al., Reference Jeong, Shim, Lee, Kim and Koh1999; Maneiro et al., Reference Maneiro, Frangópulos, Guisande, Fernández, Reguera and Riveiro2000; Jakobsen et al., Reference Jakobsen, Hyatt and Buskey2001; Stoecker et al., Reference Stoecker, Parrow, Burkholder and Glasgow2002; Gransden & Lewitus, Reference Gransden and Lewitus2003; Rosetta & McManus, Reference Rosetta and McManus2003; Kamiyama & Matsuyama, Reference Kamiyama and Matsuyama2005; Setälä et al., Reference Setälä, Autio and Kuosa2005) are known to ingest dinoflagellates or other bloom-forming microalgae, and may be important in controlling their blooms (Montagnes & Lessard, Reference Montagnes and Lessard1999; Tillmann, Reference Tillmann2004). In other cases, toxic dinoflagellates or other microalgae appear to have deleterious effects on ciliates such as changes in swimming behaviour, reduced ingestion, inability to support growth or even causing mortality (Jakobsen et al., Reference Jakobsen, Hyatt and Buskey2001; Kamiyama & Arima, Reference Kamiyama and Arima2001; Granéli & Johansson, Reference Granéli and Johansson2003; Rosetta & McManus, Reference Rosetta and McManus2003; Clough & Strom, Reference Clough and Strom2005). Moreover, several mixotrophic dinoflagellates ingesting ciliates have been described (Jacobson & Anderson, Reference Jacobson and Anderson1996; Li et al., Reference Li, Stoecker, Coats and Adam1996). For instance, Ceratium furca preys mainly on choreotrich ciliates (Smalley et al., Reference Smalley, Coats and Adams1999; Smalley & Coats, Reference Smalley and Coats2002). Thus, a predatory ciliate may become the prey of the dinoflagellate it tried to consume (Tillmann, Reference Tillmann2004).
In our study, a total of 64 ciliate species and 100 diatoms and dinoflagellates were identified during an annual cycle in Jiaozhou Bay from June 2007 to May 2008. The data are basically consistent with historical reports (Shen, Reference Shen2001; Zhang & Wang, Reference Zhang and Wang2001) in this area. 14 ciliates and 18 microalga species, which in combination successfully described their own assemblages respectively, were sought out by multivariate analyses. Furthermore, the annual variations in ciliated and microalgal assemblages all presented a clear temporal pattern and a definitely significant correlation between these two assemblage structures has been proved by the Mantel test (P = 0.001). This seasonal variation pattern in microbial community was also in agreement with the previous studies in other regions (Gomez & Gorsky, Reference Gomez and Gorsky2003; Kchaou et al., Reference Kchaou, Elloumi, Drira, Hamza, Ayadi, Bouain and Aleya2009).
In addition, our study indicated that the community structures of microplankton were different between the four seasons of a whole 1-year period and complex interspecies correlations are presented among the ciliated and microalgal species. Furthermore, in each season, planktonic ciliates and microalgae exhibited their special functions on structuring community pattern in the microbial ecosystem during each month or sample. And it is impressive that the microalgal blooms were obviously suppressed or shortened by planktonic ciliates, which has been revealed by many previous investigations (Admiraal & Venekamp, Reference Admiraal and Venekamp1986; Bochstahler & Coats, Reference Bochstahler and Coats1993; Jeong et al., Reference Jeong, Shim, Lee, Kim and Koh1999; Kamiyama & Arima, Reference Kamiyama and Arima2001; Tillmann, Reference Tillmann2004).
Moreover, 32 microplankton species represented a clear succession process and formed a circulation. Although the complexity of microbial ecosystems is enormous, with hundreds of species interacting in a number of ways from competition and predation to facilitation and mutualism (Montagnes & Lessard, Reference Montagnes and Lessard1999; Maneiro et al., Reference Maneiro, Frangópulos, Guisande, Fernández, Reguera and Riveiro2000; Tillmann, Reference Tillmann2004; Sapp et al., Reference Sapp, Schwaderer, Wiltshire, Hoppe, Gerdts and Wichels2007). Based on our data, in this microbial loop especially in each season, the specific interspecies succession process revealed the potential relationships between ciliates and microalgal species.
Numerous previous studies have documented the species-specific correlations by laboratory or field studies (Maneiro et al., Reference Maneiro, Frangópulos, Guisande, Fernández, Reguera and Riveiro2000; Jakobsen et al., Reference Jakobsen, Hyatt and Buskey2001; Stoecker et al., Reference Stoecker, Parrow, Burkholder and Glasgow2002; Gransden & Lewitus, Reference Gransden and Lewitus2003; Rosetta & McManus, Reference Rosetta and McManus2003) but it should be addressed that especially in field studies, the connections between the ciliates and microalgae cannot be proven because of a lack of a statistical resolution. In the present study, the approach of multivariate analyses basically indicates the potential relationships between dominant ciliates and dominant microalgae. Then, linear regression finally calculated the relationships and formed an interspecies correlation network. In this microbial web, it is clear that in field studies the interspecies correlations were very complex and cannot simply be defined by several species. The results showed that most ciliates and microalgae were correlated with two or more species and only in fewer cases one species was associated with another one. In a lot of previous laboratory studies, which is consistent with our results, the interspecies correlation hypotheses were examined by processing the different types of ingested matter. For example, in the study of Montagnes et al. (Reference Montagnes, Berger and Taylor1996), the planktonic ciliate Strombidinopsis cheshiri was proved to feed on diatoms and in our study S. cheshiri was definitely related to the three diatom species Guinardia delicatula, Ditylum sol and Rhizosolenia setigera. Moreover, in the studies of Smalley et al. (Reference Smalley, Coats and Adams1999) and Smalley & Coats (Reference Smalley and Coats2002), the dinoflagellate Cerarium furca was determined grazing on choreotrich ciliates, which is also proven in our study with the connections among C. furca, Totonia antarctica and Pseudotontonia cornuta. So, the microbial interspecies correlation web calculated by the linear regression between ciliates and microalgal species could be used as a robust guide for future studies especially after more and more information is gathered to verify this conclusion by further investigations on a range of marine habitats and over extended time periods. Furthermore, the Mantel and BIOENV analyses demonstrated that the temporal variations in microplankton community structures were significantly correlated with certain environmental variables, especially nutrients in combination with temperature. These findings suggest that the microplankton communities accurately reflect the water quality and have the potential for use in marine water monitoring. Moreover, the evidence supplied by multivariate analyses could guide the designs of culture researches in future. So, to discover information about complex ecological systems efficiently, this multivariate tool could be used as a powerful approach for inferring the interspecies correlations from field data in marine ecosystems.
In summary, the results of this study demonstrated that: (1) species distributions of planktonic ciliates and microalgae were both temporal in a 1-year cycle and the significant relationship between these two assemblages was represented; (2) complex interspecies correlations between planktonic ciliates and microalgae were summarized as a loop and proved by statistical evidence to form a microbial correlation web; (3) the temporal pattern of microplankton communities significantly related to the temporal changes of environmental variables; and (4) these findings suggest there is a clear temporal cycle of the microplankton communities in response to environmental changes in Jiaozhou Bay and provide basic referenced data for future field and laboratory studies and these multivariate methods have the potential to contribute a novel important tool for gaining deeper insight into the structure and stability of the microbial food web in marine ecosystems.
ACKNOWLEDGEMENTS
This work was supported by the Darwin Initiative Programme (Project No. 14-015) which is funded by the UK Department for Environment, Food and Rural Affairs, ‘The Natural Science Foundation of China’ (project number: 41076089) and a grant by the Research Group (Project No. RGP-VPP-083), King Saud University Deanship of Scientific Research. We thank Professor Weibo Song, Laboratory of Protozoology, Ocean University of China (OUC), China, for his helpful discussions.
