Tundzha (Tunca) River is the biggest segment of Maritza River who reaches Aegean Sea at Saros Bay which is classified as B class wetland according to Ramsar Pact. The river enters Turkey in Suakacağı Village of Lala- paşa-Edirne, after travelling about 350 km throughout Bulgarian land. The length of Turkey segment of the river is 54.7 km and the mean flow rate is 18.38 m ³ /sec. The river meets with Maritza River to the south of Edirne city centre. The areas surrounding the river, as in Thrace region in general are used for corn, sun flower, crop, fruit, vegetable and particularly rice cultivation. Local woodland and scrubby areas along the river also exist. A preliminary study was performed and 5 stations were determined to represent the part of the river in Turkey from the Bulgarian border until it meets Maritza River (Figure 1). Water and phytoplankton samples were taken from the 5 stations twice a week from June 2002 to October 2003 and once a month from November 2002 to May 2003.

A Ruttner water sampler was used to obtain water samples just below the surface of the water body in order to determine some physico-chemical properties of the river. Water temperature, Dissolved Oksigene, pH and con-

ductivity were measured on site during samplings using field type equipments (Lovibond sensodirect oxi 200, pH 200 and Con 200) and PO₄-P, NO₃-N, NO₂-N, total hardness, COD and BOD were measured in laboratory in accordance with APHA-AWWA-WPCF methods [14] . Chlorophyll-a was analyzed according to Nusch [15] . Winkler method was applied in dissolved oxygen determinations [16] . BOD and COD values were measured using methods given in Golterman [17] .

Sampled phytoplankton specimens were identified by investigation of temporary preparations. For this purpose, water samples were filtered from Whatman GF /A ( 90 mm diam., 1.6 µm pore size) paper with the help of a water trompe and dissolved in 10% glycerine. An Uthermol counting chamber was used to calculate the organism number per liter [18] [19] . Biovolumes were measured using geometric or combined forms formulae according to Standard Operating Procedure for Phytoplankton Analysis [20] . For the identification of algal species Husted , Cleve-Euler, Pestalozzi , Prescott , Komarek and Fott, Krammer and Lange-Bertalot were used [21] - [29] . And also all the species are checked in algaebase cite [30] .

A Bray-Curtis analysis was performed to reveal similarities, if any, among stations based on algal species diversity and abundance [31] . Abiotic variables were correlated with main phytoplankton attributes using non-para- metric Spearman’s correlation coefficients. Differences in all variables were tested using a non-parametric Kruskal-Wallis ANOVA median test (KW). Differences at the <0.05 level were accepted as significant. Also, to compare stations, non-parametric statistics (the Mann-Whitney U test) were used. Species richness was considered to be the number of taxa present in each sample. Biological diversity (H') was calculated by the Shannon and Weaver [32] .

The results of Physico-chemical analysis showed that the highest pH and dissolved oxygen values were measured in Station 1 in spring and winter months, respectively. The Cl content appeared to be high in station III in spring and Biological Oxygen Demand in station V in winter. Station V was also characterized with the highest Phosphate value in summer months. NO₂-N and NO₃-N was measured at high levels in stations III (in auntumn), respectively. The results also revealed that Chlorphyll_a was measured in its highest value in station IV in summer. The mean results of all analyis are given in Table 1.

Nitrate-N was found to be negatively correlated with water temperature, electrical conductivity and biomass and no significant relation was found between other physico-chemical variables (P < 0.05). The comparison of stations based on abiotic factors revealed that all stations were similar (P < 0.01).

The taxonomic evaluation of all samples obtained showed that a total of 138 taxa lived in the river (Table 2). Station I was the richest in terms of species diversity with 118 taxa while station 5 was the least diverse station with 87 taxa (Table 3). Chlorophyceae was found to be represented with the highest number of species (65 taxa) followed by Bacillariophyceae (46 taxa), Euglenophyta (13 taxa) and Cyanophyta and Charophyta (7 taxa). Since none of the 7 taxa of cyanophytes showed an extreme growth during the study period, they never were among the dominating or subdominating taxa and thus did not lead to formation of water bloom.

Chlorophyta, which was represented with 57 taxa in station I (richest) and 39 taxa in station V (poorest), appeared to be the dominant group in the phytoplankton community in the river. On the other hand, diatoms constituted the second dominant group and all stations were similar in terms of their diatom compositions. Although pennant diatoms were rich in terms of species diversity, Cyclotella species of centric diatoms were found to have the highest cell counts in all stations, particularly in spring, summer and autumn months.

The highest cell counts were obtained in the sample from 1th station (3,459,313 cells∙L⁻¹) whereas the lowest counts were obtained from samples of 2th (1,713,312 cells∙L⁻¹) and 5th stations (1,815,672 cells∙L⁻¹). Centric diatoms were highly abundant in the samples from the river (Figure 2). Cyclotella species, in particular, were

obtained with high biomass values in all stations by the beginning of spring season. Diatoms appeared to be the dominant group in terms of cell counts followed by green algae as the second dominant group in the river. Other algal groups were recorded with low abundance values (Table 4). Most of the algal species found in the Bulgarian segment of the river were also determined in the present study except Ceratium, Peidinium, Dinobriyon and Stephanodiscus [13] .

Differences in the biovolume of phytoplankton from the 5 sampling sites are illustrated in Figure 3. Diatoms are the prevailing group in samples from all localities except the 4th station where Chlorophyta was the dominant group. The biovolume of Euglenophyta members were measured to be highest in the 4th station (2028.8 µg∙L⁻¹) whereas it was the lowest in the 5th station (1366.8 µg∙L⁻¹).

When floristic compositions of all 5 stations were used to obtain a similarity index through the cluster analysis, the 2th and 3rd stations were found to have very similar floristic compositions and, although all stations revealed a somewhat same similarity, the 5th station was grouped to be the outgroup among the stations with the least similarity (Figure 4(a)). The cluster analysis based on biovolumes measured for each stations revealed that the 1th and 3th stations were the most similar ones to each other (Figure 4(b)).

The abundance of phytoplankton was significant, and was positively correlated with Chl-a and temperature (r = 0.77 and r = 0.86, respectively P < 0.01). The biovolume of Bacillariophyta was positively correlated with the species number (r = 0.71, P < 0.01). The estimates of algal biomass did not show any correlation with nutrients. Some water quality parameters were found to show a significant relationship with the dominant taxa. For example, DO concentration was correlated with the mean total number of C. meneghiniana and M. varians (r = 0.72

and r = 0.61, respectively P < 0.05). In addition, nitrate and phosphate were found to negatively correlated with biovolume (r = −0.71 and r = −0.54, respectively P < 0.05).