The Horizontal Distribution of Siliceous Planktonic Radiolarian Community in the Eastern Indian Ocean

The plankton radiolarian community was investigated in the spring season during the two-month cruise ‘Shiyan1’ (10 April–13 May 2014) in the Eastern Indian Ocean. This is the first comprehensive plankton tow study to be carried out from 44 sampling stations across the entire area (80.00◦–96.10◦ E, 10.08◦ N–6.00◦ S) of the Eastern Indian Ocean. The plankton tow samples were collected from a vertical haul from a depth 200 m to the surface. During the cruise, conductivity–temperature–depth (CTD) measurements were taken of temperature, salinity and chlorophyll a from the surface to 200 m depth. Shannon–Wiener’s diversity index (H’) and the dominance index (Y) were used to analyze community structure. There was a total of 168 plankton species, composed of Acantharia, Phaeodaria, Polycystina, Collodaria and Taxopodida (monospecific—Sticholonche zanclea, Hertwig is the only recognized species). Hence, it included both celestine-based and siliceous organisms, which are also described here for the first time from this region. Total radiolarians ranged from 5 to 5500 ind/m−3, dominated by co-occurrences of Sphaerozoum punctatum and Stichonche zanclea species at the south-equator zone (SEQ)-transect 80◦ E and equator zone (EQ)-transect Lati-0. The possible environmental variables were tested through RDA analysis; although no result was obtained for the full species dataset, the samples from the equatorial transect related strongly to mixed-layer chlorophyll a concentration and those of a north–south transect to surface silicate concentrations or mixed-layer nitrate were significantly correlated (p < 0.01) to the radiolarian community. Our results indicate that the silicate and chlorophyll-a concentrations are the two major factors affecting the radiolarian distribution along two of the investigated transects (southern equator and equator) in the study area.


Introduction
Radiolaria are the siliceous-shelled, single-cell, planktonic organisms frequently studied for palaeoceanographic and environmental reconstruction. They are globally distributed and widely    Only a few studies have been conducted in the Eastern Indian Ocean and most of them are based on surface sediment samples and the investigation of polycystine radiolarians [24,25,27]. This is the first comprehensive plankton tow sampling study in the Eastern Indian Ocean. The research objectives were to estimate the diversity and horizontal distribution of the radiolarian community

Sampling Sites and Collections
Plankton-net tow sampling was conducted in the Eastern Indian Ocean (80.00 • -96.10 • E, 10.08 • N-6.00 • S). The study area covered the ca. 1000-1500 km distance from the Bay of Bengal (BOB) and parallel to the Sumatran coast of Indonesia to equatorial section and southern equatorial zone ( Figure 2). The first transect (hereafter "Long-90") consisting of the 17 stations along the NNW line from 5.5 • S, 96.1 • E (station I511) to 10.08 • N 88.1 • E (station IQ2), the second (hereafter, "Long-80") of 15 stations along two adjacent north-south lines from 2.42 • N, 80.0 • E (station I302) to 6.0 • S 80.0 • E (station I318) and the third, which partially overlapped the other two 11 stations (hereafter "Lat-0"), closely followed the equator (Table 1).
Using a modified Indian Ocean Standard Net (20 µm mesh size, 0.57-m diameter, 470 cm, with a mechanical flow meter), samples were collected from a vertical haul from a depth 200 m to the surface. About 1000 mL of sample was collected in each PM bottle and fixed with 2% formalin preservative. Preserved samples in the PM bottles were returned to Dr. Sun's laboratory (Tianjin University of Marine Science and Technology, China) for analyses.

CTD Data Collection
Temperature and salinity data were collected using a Seabird conductivity-temperature-depth (CTD) system (SeaBird SBE 911 plus V2). Using the Strickland and Parson [36] method, a standard protocol to measure the chlorophyll a, was employed using 500 mL of seawater filtered through a 25-mm diameter Whatman filter paper and kept in 5 mL of acetone (93%) at 20 • C for 24 h, then analyzed using a fluorometer (Trilogy, Turner Designs, Sunnyvale, CA) for estimation of chlorophyll a.

Nutrients Concentration
The nutrients, silicate, nitrate, phosphate and dissolved oxygen values were derived from the interpolation of the objectively-analyzed values during boreal Spring data (April to June) collected by World Ocean Atlas 2013 (https://www.nodc.noaa.gov/OC5/woa13/woa13data.html) at the 25

Community Analysis and Microscopic Observation
To enumerate the community structure and for primary identification, fixed samples were examined at 20× and 40× magnification under an inverted light microscope (Motic, AE 2000, Xiamen, China) which was equipped with a digital Camera (Moticam 2506, 5.0 m, Pixel, Xiamen, China). About 2 mL of each sample were examined to count cells in a settling chamber using a modified Utermöhl method [37]. Species identification was confirmed by the use of scanning electron microscopy (SEM). Samples were pre-screened through a net again to remove large zooplanktons and subsamples were filtered on 0.6 µm, pore-size Millipore filters and air-dried for 24 h, then mounted on a stub, sputter-coated and examined under the SEM microscope (JEOL JSM-IT300LV, Hitachi, Japan) at Biological building (13), Tianjin University of science and Technology, TEDA, China. Species were identified using published taxonomic references [13,[38][39][40].

Data Analysis and Statistical Methods
Shannon-Wiener diversity index (H ), Pielou evenness index (J) and Dominance index (Y) has been used by the modified equation [37] to analyze the composition and richness.
where S is the phytoplankton species of each sample [41]. N is the total number of cells of all species counted; ni is the cell number of species i; Pi = ni/N represents the relative abundance of a given species; fi is the frequency of occurrence of species i in each sample. Species with Y > 0.02 was defined as dominant species [42].

Correspondence Statistical Analysis
The Hellinger-transformed Legendre and Gallagher [43] of species data were investigated by redundancy analysis ("RDA") using the R software [44] RDA was chosen to allow the use of ordiR2step (Blanchet et al. [45] and an equivalent of Canoco (ter Braak and Šmilauer, [46] for the coefficient freedom value plus Pearson correlation values)).

Environmental Dataset of Spring 2014 in Eastern Indian Ocean
An environmental dataset was constructed comprising the values for temperature, salinity and chlorophyll a measured during the cruise at the depths (0 m, 50 m, 75 m, 100 m and 200 m) were derived from CTD data (Figure 4a-i).

Vertical and Horizontal Profile of Temperature and Salinity
The vertical distribution of temperature and salinity in the three transects are summarized in Figure 4). The temperature and salinity in the depth 0-50 m ranged from 12.82 and 32.55 psu and in the depth 100-200 m ranged from 31.82 • C and 32.5 to 35.5 psu, respectively ( Figure 4). Horizontally, maximum temperature value (>31.82 • C) was observed in the BOB samples (HFA01, HFA05, I601, I603, IQ2) and (<30 • C) was observed in the equator samples (I401-I413) (Figure 5a). The maximum salinity was observed also in the equator stations (I308, I401, I403, I406) at the transect 80 • E and minimum value of salinity (>32 psu) was observed in BOB sampling areas (IQ2, HFA01-HFA05) (Figure 5b). The low salinity values in the BOB could be influenced by riverine runoff inputs and aggravate stratification. However, the salinity remains steady in the area nearby the Sumatran coast of Indonesia (Figures 4b and 5b), which probably contributed to an increase in salinity gradients due to riverine influences from these coastal locales.

Vertical and Horizontal Profile of Temperature and Salinity
The vertical distribution of temperature and salinity in the three transects are summarized in Figure 4). The temperature and salinity in the depth 0-50 m ranged from 12.82 and 32.55 psu and in the depth 100-200 m ranged from 31.82 °C and 32.5 to 35.5 psu, respectively ( Figure 4). Horizontally, maximum temperature value (>31.82 °C) was observed in the BOB samples (HFA01, HFA05, I601, I603, IQ2) and (<30 °C) was observed in the equator samples (I401-I413) (Figure 5a). The maximum salinity was observed also in the equator stations (I308, I401, I403, I406) at the transect 80° E and minimum value of salinity (>32 psu) was observed in BOB sampling areas (IQ2, HFA01-HFA05) ( Figure 5b). The low salinity values in the BOB could be influenced by riverine runoff inputs and aggravate stratification. However, the salinity remains steady in the area nearby the Sumatran coast of Indonesia (Figures 4b, 5b), which probably contributed to an increase in salinity gradients due to riverine influences from these coastal locales.

Vertical and Horizontal Profile of Temperature and Salinity
The vertical distribution of temperature and salinity in the three transects are summarized in Figure 4). The temperature and salinity in the depth 0-50 m ranged from 12.82 and 32.55 psu and in the depth 100-200 m ranged from 31.82 °C and 32.5 to 35.5 psu, respectively ( Figure 4). Horizontally, maximum temperature value (>31.82 °C) was observed in the BOB samples (HFA01, HFA05, I601, I603, IQ2) and (<30 °C) was observed in the equator samples (I401-I413) ( Figure 5a). The maximum salinity was observed also in the equator stations (I308, I401, I403, I406) at the transect 80° E and minimum value of salinity (>32 psu) was observed in BOB sampling areas (IQ2, HFA01-HFA05) ( Figure 5b). The low salinity values in the BOB could be influenced by riverine runoff inputs and aggravate stratification. However, the salinity remains steady in the area nearby the Sumatran coast of Indonesia (Figures 4b, 5b), which probably contributed to an increase in salinity gradients due to riverine influences from these coastal locales.

Horizontal Distribution of Radiolarian in EIO
The abundance, diversity and taxonomic patterns within the three transect zone show the following characteristics.
1. NEQ transect-90° E: Radiolarian abundance ranged from 6 to 1080 ind/m −3 which recorded from 17 stations at transect 90° E (Figure 12c). These samples contained a total of 89 species. Each species has heterogeneous distribution, 2-23 species were found from each station and maximum assemblage occurred at the stations (I609) (Figure 12b). The maximum abundance was recorded at the BOB stations (AH-05), followed by 661 ind/m −3 at stations (I109, I609) which were dominated by the occurrences of Stylodictya multispina, C. inerme and Siphonosphaera polysiphonia ( Figure 13; Table 2) and low abundance <250 ind/m −3 was recorded by the

Horizontal Distribution of Radiolarian in EIO
The abundance, diversity and taxonomic patterns within the three transect zone show the following characteristics.

2.
Equatorial transect-0 • : The abundance of radiolarian in the equatorial samples area ranged from 14 to 996 ind/m −3 and collected from 11 stations (Figure 12c). These samples contain total of (39 species) and approximately, 2-15 species were found from each station (Figure 12b). The diversity increases maximum of 15 species at stations (I403, I412, I413) but subsequently drops to two species at other stations. The assemblage in the equatorial region is dominated by S. punctatum, S. cf ovodimare, L. pentagona pentagone and S. zanclea species (Figure 13).  2. Equatorial transect-0°: The abundance of radiolarian in the equatorial samples area ranged from 14 to 996 ind/m −3 and collected from 11 stations (Figure 12c). These samples contain total of (39 species) and approximately, 2-15 species were found from each station (Figure 12b). The diversity increases maximum of 15 species at stations (I403, I412, I413) but subsequently drops to two species at other stations. The assemblage in the equatorial region is dominated by S. punctatum, S. cf ovodimare, L. pentagona pentagone and S. zanclea species (Figure 13).

Identification of Potentially Explanatory Environmental Variables
The RDA-generated inertia and the first ten eigenvalues appear as Table 3. Table 3. The inertia and first ten eigenvalues were generated by RDA from the full species data and the three transect (see text). Two sets of data are given for each source: That for the full, Hellinger-transformed, species set and that for the reduced to include the dominant species only. Two tests were employed to determine how many, if any, variables the eigenvalues from the species dataset might explain. The first test used the Guttman-Kaiser Criterion ("GKC") (Yeomans and Golder [47]): Fourteen eigenvalues exceeded the mean eigenvalue, although two only marginally, suggesting many explanatory variables. The second, more stringent, test employed the "Broken Stick" Model ("BSM") Jackson [48] and this indicated that there might be about two explanatory variable. The test results are plotted as Figure 14a.

Source
An environmental dataset was constructed comprising the values for temperature, salinity and chlorophyll measured during the cruise at the ocean's surface and depths of 25, 50, 75, 100, 150 and 200 m below sea level (bsl) plus the World Ocean Atlas 2013 objectively-analysed values for nitrate, phosphate, silicate and dissolved oxygen for the season April-June at the same depths [49,50]. RDA was applied to the species data and each variable in the environment dataset in turn and Analysis of Variance (ANOVA) run on the output: Those environmental variables where the probability of the F statistic was greater than 0.005 and the ratio of the first RDA axis to the first unconstrained axis was greater than one was to be published. No significant results were found. Because the species data were drawn from a relatively large ocean region with a range of environmental conditions, three transects were selected to examine whether plankton abundances were affected by different environmental variables in different parts of the study area ( Figure 2 and Table 1). An environmental dataset was constructed comprising the values for temperature, salinity and chlorophyll measured during the cruise at the ocean's surface and depths of 25, 50, 75, 100, 150 and 200 m below sea level (bsl) plus the World Ocean Atlas 2013 objectively-analysed values for nitrate, phosphate, silicate and dissolved oxygen for the season April-June at the same depths [49,50]. RDA was applied to the species data and each variable in the environment dataset in turn and Analysis of Variance (ANOVA) run on the output: Those environmental variables where the probability of the F statistic was greater than 0.005 and the ratio of the first RDA axis to the first unconstrained axis was greater than one was to be published. No significant results were found. Because the species data were drawn from a relatively large ocean region with a range of environmental conditions, three Application of the GKC to the transect species data suggested that Long-90 and Long-80 might each have about four explanatory variables and Lat-0 three. The BSM indicated that no explanatory variable would be revealed. RDA and ANOVA were applied to each of the transects with the same publication criteria as for the full species dataset. Transect Long-80 gave five and Lat-0 gave two results significant at the 0.1% level; Long-90 yields no significant results-See Table 4. An R program was written to use ordiR2step to find the most likely additional explanatory environmental variables given the five already found for transect Long-80 and the two for Lat-0. The program's results suggested some secondary explanatory variables but, when strongly correlated pairs (see Table 5) were eliminated, none remained. The most likely explanatory environmental variables (based on the Pr(>F) statistic) are, therefore, for the Long-80 transect is silicate and for Lat-0 chlorophyll a, both at the sea surface.

Discussion
There are only a few sediment sampling studies of radiolaria have been conducted in the Eastern Indian Ocean so far [24][25][26]. We describe a comprehensive study of the planktonic radiolarian community and its horizontal distribution on the three transects in the Eastern Indian Ocean. We made no distinction between living and dead radiolarian specimens, the samples were immediately fixed with formalin after collection and no evidence was gathered to determine how many were living at the time of collection, therefore our samples represent both the living and dead individuals that were in the water column from sea surface to 200 m depth at the time of plankton net sampling.

Radiolarian Communities in the Eastern Indian Ocean
Preliminary studies reported only the 35 taxa from the Western Indian Ocean [24], 74-261 taxa from the Eastern Indian Ocean [25,27], 40 taxa from the BOB [26]. In this study about 168 taxa of the major radiolaria group, such as Polycystinea, Acantharia, Collodaria, Taxopodida and Phaeodaria were collected from plankton tow samples from the surface to euphotic zone of EIO (Figures 6-11).
Most of these groups were tropical or sub-tropical Pacific and Atlantic Ocean dwelling species [3,16]. On the base of previous geographical records, the distribution of radiolarian species in the EIO does not reflect their Pacific and Atlantic counterparts, was described here for the first time in the EIO (Table 7). However, the species of Polycystinea group were also retrieved, which were occasionally important (Supplemental Table S1; Figures 8 and 9). The Collodaria and Taxopodida group were mainly represented by the high frequency distribution 70-90% which were dominant in our samples (Figure 12d,e), have a widespread distribution and contribute high fluxes within phytoplankton communities in open waters [51]. Nonetheless, this is the first report on the co-occurrence of Collodaria and Taxopodida in the Eastern Indian Ocean. However, variations of Collodarian contribution to the radiolarians were lower in the Antarctic and Indian Ocean [51]. The species of Spumellaria (15%) and Nassellaria (6%) were found in relatively low abundance in all samples (Figure 12g,h). The Spumellaria and Nassellaria species were highly abundant in the BOB [26] and in the Indian Ocean [1]. Spumellaria and Nassellaria have a global distribution in the surface and sediments waters of the Atlantic and Pacific Ocean [1]. More recently, a high percentage of Spumellaria were correlated with a low percentage of Collodaria and Nassellaria in the South China Sea [52]. We additionally found the Acantharia and Phaeodarian in low numbers in our samples, contributing only about 8-13% to the total radiolarian population (Figure 12f,i) and notably occurred in the twilight zones of the Ocean [51]. Acantharia are usually found in the tropical region of the Atlantic Ocean and East China Sea (Table 7) and Phaeodarian are predominantly found in the Adriatic Sea and Arctic Sea [53]. However, Gupta et al. [26] reported the few Phaeodarian species from sediments of BOB. Despite their low abundances, they are a significant component because they are celestine and opaline silica protists. We can conclude that more work is needed to examine the sampling depth per station and to distinguish between living and dead radiolaria in the Eastern Indian Ocean.

Characteristics of Radiolarian Abundance, Horizontal Distribution and Environmental Factor
Diversity indices and abundance was higher around the equator and SEQ transect regions ( Figure 12). One of the reasons for high diversity was the occurrence of two species, S. punctatum and S. zanclea, in these two transect area. The diversity of Collodaria species such as S. punctatum and S. polysiphonia can be influenced by the trophic status of oceanic provinces and increased towards more oligotrophic regions [51]. As a result, high abundance in the equatorial region at stations (I815, I310, I312) can be predicted by the Wyrtki jets in spring season when WJs carries the high salinity and high nutrient waters from west to eastward [34]. However, low diversity was observed at the northern BOB and near Sumatran coastal area due to less saline waters due to runoff from the Ganges river and southwest upwelling region from these coastal locales. Radiolaria are rare in coastal areas, bays, estuarine and fjord areas due to less saline water [21,25]. This result is contrary to recent research in the Pacific Ocean, the radiolarian abundance and distribution in the continental shelf area of the Pacific Ocean was reported under low salinity and high temperature [66].
Radiolarian abundance in the Eastern Indian Ocean during this study was ranged from 5 to 5850 ind/m −3 ( Figure 12). The high abundance at the one station (I815) from the south equator transect, where the colonies of S. punctatum and S. fuscum were very abundant (Figure 6d,f). Over all, abundance was below the 1500 ind/m −3 . Our results generally agree with previous information that radiolarian densities are higher in the shallow water [1,16]. Plankton net-studies in the Pacific and Atlantic Ocean also show that the highest abundance of radiolarian occurs in the upper ocean layer [21]. The abundance of radiolarians was similar to the other studies from the Atlantic Ocean [13] and Indian Ocean [1]. It ranged from 1 to 100 m −3 and 500 m −3 and polycytinea radiolarian numbers arẽ 1200 times greater at 0-150 m than at 500-1000 m [13]. At the depths (100 m), the maximum densities such as (1400 specimen/m −3 ) and low densities (250-500 ind/m −3 ) can be observed and reported from the Pacific Ocean [4,16]. Boltovoskoy and Correa, [16] also recorded the maximum abundance such as (~200,000 specimens/m  Figure 13c) and are restricted to shallow photic layers of the ocean [5].

Factors Influencing on the Radiolarian Distribution
The vertical water structure of 0-50 m has the characteristic of high temperature and low salinity and high nutrients. At the depth below 100-200 m, it is characterized by high salinity and low temperature as shown in Figure 4. BOB is located in the north which has the characteristics of intrusive river runoff with low salinity and high temperature in the EIO [31], Sumatra is open coastal zone with conjunction of the BOB, characterized by low salinity and oxygenated water and the equator area is characterized by warm water and mixed layer of Chlorophyll a [30].
In our study area, the temperature and salinity of the equator are high and the boreal-spring nutrient (WOA13) data shows that the equatorial-nutrient concentration is low (Figure 3e-g). The RDA analysis gave most of the negative but significant correlation values between the environmental variables ( Table 5). RDA explains the positive correlation at level (p < 0.05) between the Chlorophyll a, nitrate and temperature and between the salinity, silicate and phosphate (Table 5). Dissolved Oxygen was negatively correlated, indicating the minimum OMZ condition. This suggests that more than one variable, e.g., Chlorophyll a, temperature and silicate can shape the radiolarian community in the EIO [69,70].

Factor Influence on Radiolarian in Three Transect Area, EIO
Salinity and temperature are two significant factors that are closely related to the radiolarian community at the depth of 200 m [25][26][27]52,70,71]. High temperature can be associated with high radiolarian diversity [52]. Rogers and De Deccker, [27] clearly demonstrate that, temperature and salinity between the 100 m and 300 m and nutrients at the 100 m in the Eastern and Southern Indian Ocean, have strongly impacted on the distribution of the radiolarian community. In the present study, the relationships between the all species dataset, within each transect, with seven environmental variables often did not provide any significant correlations (Figure 14a), but significant relationships were also found with the dominant taxa at the transect 80 • E and equator ( Figure 14). The ratio of first RDA axis to the first unconstrained axis was greater than 1 and probability of F statistic was significant 0.005, but they did not provide a significant correlation at the transect long-90 • E. The difference between the initial and subsequent unconstrained eigenvalues was not great, an attempt was made to strengthen the effect by excluding those species with an RDA score of less than two standard deviations from the mean score in any of the first six components-That is using only the dominant species. The ratio of the first to the second unconstrained eigenvalues did increase (Table 3-labelled "Dominant") but application of GKC, BSM and RDA failed to produce any better results than using all the species data. It was noted that the species data matrix was very sparse (approximately 90% of the cells contained zeros). A dense species dataset was generated by deleting the sparsest species records and the battery of tests described above applied. No significant results were obtained. The variance in the full species data cannot be explained by a small number of variables probably because the study area is large enough to cover a number of ocean environments. Analysis of two sections of the total species data revealed possible explanatory variables: (a) The variance in species from a north-south transect at approximately 80 • E ("Long-80") could be due to the surface silicate concentrations but, alternatively, might result from nitrate concentrations or dissolved oxygen in the mixed layer. Silicate, nitrate and oxygen are necessary for radiolarian growth so the connection with the variance is likely to be causal. (b) The variance in an equatorial transect ("Lat-0") is strongly related to the chlorophyll a level in the mixed layer but it is not certain that chlorophyll a is necessary for radiolarian growth. Thus, the connection may be indirect such as radiolarians may consume chlorophyll-producing organisms or may simply require a similar environment. However, neither temperature nor salinity showed significant correlations with the radiolarian distribution ( Table 4). As salinity and temperature enable radiolarians to thrive in different water masses [16]; its effect on the vertical distribution is greater than that on the horizontal distribution [23].

Factors Influencing the Radiolarian Communities
Temperature and salinity are not sufficient explanatory variables: Chlorophyll a and nutrients and dissolved organic matter (DOM) can stimulate heterotrophic communities such as the radiolarian [4,27,71,72], tintinnid [29] and diazotroph communities [73,74]. These environmental variables play an important role in circulatory changes of EIO in the spring season. The nutrients (silicate, phosphate and nitrate) may be depleted throughout the EIO, except in coastal areas, when coastal upwelling cause nutrients to elevate [31]. In this study, the Chlorophyll a concentration in the surface layer ranged from 0.005 to 0.89 µg/L, which is similar to the remote sensing Chlorophyll-a data and high chlorophyll a of the same period can be seen in this region (Figure 4; Supplemantry Figure S1). The Chlorophyll a value is significantly higher in spring season in same region [75]. This strengthens the possibility of using Chl-a remote sensing data to estimate the geographic distributions of radiolarian in the Ocean. These concentration values often lead to low nutrients in the DCM layer in spring [29,31]. Our findings also confirmed earlier reports that low nutrients at the equator correspond to the DCM (deep Chl a concentration maximum) layer from 50 to 100 m in a same study area [29]. This DCM chlorophyll a maximum has caused the dominance of nano and picoplankton biomass [29,31] including some heterotrophs bacteria [73] in the region. The silicate, nitrate or oxygen concentrations including chlorophyll a in surface water are mostly highly correlated at the transects of 80 • E and equator (Table 4). Among environmental variables, the only significant positive correlation was observed between the S. punctatum and A. cineraria with temperature and L. dodecantha with Chl a concentration (p < 0.005) ( Table 6). Differently, a negative correlation was observed for S. polysiphonia with phosphate at a high significant value (p < 0.001) and salinity, D. tetrathalamus and H. armatum-hostile, with silicate concentration ( Table 6). The significant correlation between radiolarian species with SST and SSS also has been reported from the Pacific Ocean [66].
These were the major environmental factors affecting on the distribution of silicified dominant species including Collodarian in the EIO (Figure 15). This finding supports previous research in Australia [76], which suggested that the symbionts of such protists can take up nutrients from their hosts, so nutrients affecting their distribution in that area and so their presence at low latitudes and absence from high ones are to be expected. The symbiotic species of Collodarian were dominant in the study area, which is associated with high radiolarian densities in numerous studies of the tropical ocean [2,5]. The symbiotic radiolarians are key player for the carbon export from nutrient-poor and subtropical marine areas [2,10]. Recently, radiolarian is contributing 29% to the meso-zooplankton in the ocean, cause the high biomass 0.089 pg C in the upper column 200 m depth and about 5.2% of total carbon biomass [2]. Photosymbiotic Collodarian species often found with the high abundant at the 100 m depths and with Acantharia contributed only 0.18% primary production in the oligotrophic waters [2]. However, Polycystinea radiolarian at the 200-400 m constituted of high biomass to the total Rhizarian [4]. The energy obtained by their primary production can enhance carbon uptake and increase carbon fluxes due to the high rates of primary production (1400-41,000 ng carbohydrate colony −1 h −1 ) and result in high carbon fixation (91.16 µg) [77,78]. Earlier studies indicated that colonial radiolarians in the subsurface have a strong correlation with chlorophyll a maximum at the subsurface layers (25-75 m) and cause high fluxes in the Gulf of Mexico [76] and also in the Pacific Ocean [4].
Our studies suggest that gelatinous packed species belonging to Collodarian were abundant in the euphotic zone and dominate the radiolarian community. Many species occurring in the surface areas are particularly herbivorous and omnivorous planktons which can be indirectly related to chlorophyll a, similar to other studies [72,79]. Recently, bacteria [28], tintinnids [29], coccolithophores [30], phytoplankton studies [31] have been conducted at the same sampling area and these can be nutrient source of radiolarian [23]. The radiolarian depicting in the surface for feeding the chlorophyll a producing organism such as bacterial feeders or crustaceans or it is possible that the species may be required similar environment.

Conclusions
The results presented here describe the first comprehensive plankton-net sampling of the Eastern Indian Ocean. Planktonic radiolarian within five classes, e.g., Acantharia, Taxopodida, Collodaria, Phaeodaria and Polycystinea, including 60 newly reported taxa, were described in this study area. Community composition, abundance and distribution of radiolaria are different in each transect but, in each case, are dominated by the combined occurrences of Collodaria and Taxopodida.
The highest abundance was recorded in the south-equator region, and the horizontal distribution appears to be controlled by three explanatory variables; chlorophyll a, silicate, and nitrate dominate the radiolarian species: (a) Based on the RDA analysis, the radiolarian abundance in the north-south transect at approximately 80 • E ("Long-80") is mainly controlled by the surface silicate concentrations or might be possible for high nitrate concentrations or dissolved oxygen in the mixed layer. So, their connection with the growth of radiolarian is casual and dominates only the S. punctatum, S. fuscum and S. zanclea species. (b) Radiolarian abundance in an equatorial transect ("Lat-0") is strongly related to the chlorophyll a level in the mixed layer which dominates the S. punctatum, L. pentagona pentagone and S. zanclea but it is not certain that chlorophyll a is necessary for radiolarian growth. Thus, the connection may be indirect, such as radiolarians may consume chlorophyll-producing organisms or may simply require a similar environment.