METHODS

Water sampling
Eleven lakes were monitored during February and April 2009 corresponding to lean monsoon season with only sewage inflow and higher evaporation. Three replicates of water samples were collected from inlets, outlets and other sites (such as centre) to observe the effect of sewage and effluents as well as to understand the quality variations at the regional scale. Onsite variables like pH, electric conductivity (EC), salinity (SAL), total dissolved solids (TDS), water temperature (WT) (Extech pH/conductivity EC500) and dissolved oxygen (DO) were measured. The samples were stored at 4 ⁰C in laboratory and analyzed for nitrates (N), inorganic phosphates/or phosphorous (P), total hardness (TH), calcium hardness (CaH), magnesium hardness (MgH), chlorides (CHL), alkalinity (ALK), chemical oxygen demand (COD) and biological oxygen demand (BOD) following standard protocols (Trivedy and Goel, 1986 and APHA, 1998).

Diatom sampling, identification and analysis
Diatoms samples were collected along with the water samples from the microhabitats (i.e., Epilithic sample: 5-15 cm cobbles and pebbles) following the standard collection techniques (Taylor et al. 2007b; Karthick et al. 2010). Epiphytic samples were collected from water hyacinth (Eichornia sp.) by crushing the submerged roots of plant in a polythene cover. Epilithic samples were collected by brushing the stones thoroughly with tooth brush. Episammic sample was collected from upper layer of sediment (0.5 cm) using a pippete. All samples were transferred to polythene bottle and carried to laboratory for immediate observation in order to record live diatoms and later preserved in ethanol. Diatom samples were cleaned by KMnO₄- acid method following standard protocols (Taylor et al.2007b; Karthick et al. 2010) for further enumeration. 350-400 frustules in each slide were counted using DIC microscopy (Olympus). Diatoms were identified following taxonomic literatures (Krammer and Lange-Bertalot (1986, 1988, 1991), Lange-Bertalot, (2001), Taylor et al. 2007c and Karthick et al. 2008). Diatom taxa occurring at least in one sample with a relative abundance of 5 % were included in the analysis. Trophic diatom index (TDI) indicating the nutrient load in the sample was calculated in OMNIDIA 7.0 software.

Macroinvertebrate sampling and identification
Insects and mollusk communities as macroinvertebrates were collected during the summer (April -May 2009).  Aquatic insects were sampled using a D-net sampler holding against water current and dragged along the shore of the lakes up to a distance of 1m. Contents of the net were pooled and preserved in 70% ethanol. At each lake, mollusks in the littoral zone were collected by employing quadrat and time constrained methods. In time constrained method, mollusks were sampled by search and handpick along the littoral zone for five minutes. While in quadrat collection, aquatic vegetation and other objects were collected in an area of 25 × 25 cm from littoral zone and were kept in a bucket and half of the bucket was filled with water and all the materials were washed thoroughly. Then the bucket water were passed through 0.5 mm sieve, mollusks retained were collected. The samples collected from both methods were kept in separate glass/plastic bottles and preserved with 70 % alcohol. Collected samples were examined under a stereo zoom microscope (10X) and identified using standard taxonomic literatures (Fraser, 1933-36, Morse et al. 1994 and Subramanian and Sivaramakrishnan, 2007 for Insects and Ramakrishna and Dey (2007) and Subba Rao (1989) for mollusk).

Statistical analysis
Percentage relative abundance of species was derived along with the number of taxa, dominance (D), evenness (E) and Shannon and weaver diversity index as quantitative structural attributes. Principal component analysis (PCA) was performed to assess significant environmental variables. Spearman correlation between diatom indices (TDI, GDI) and macroinvertebrate indices (FBI, BI) with EC, DO, N, P and alkalinity was calculated to observe the significant correlations. All the statistical analysis was carried out in PAST version 2.06.