Chemical analyses and estimation of plankton biomass
In the laboratory, sestonic particles in the water were concentrated onto pre-combusted GF/F (0.7 µm mesh size) filters. Seston carbon and phosphorus were determined with a CN analyzer (model 2400; Perkin-Elmer, Inc.) and by the ascorbate-reduced molybdenum-blue method, respectively. Chlorophyll a was extracted by 90% ethanol for 24 hours in the dark and quantified using a fluorometer (TD-700; Turner Designs, Inc.). For measuring abundance of phytoplankton and small zooplankton, 250 to 500-mL of the collected water were fixed with dilute Lugol’s solution and concentrated into 20 mL by gravity. For each phytoplankton taxa, number of cells in 0.2 to 1.0 mL of the concentrated sample were counted and size of 20 to 50 cells were measured for estimating cell biovolume. For small zooplankton, we counted all individuals in the concentrated samples according to taxa with measurements of body length and width. For crustacean zooplankton, we categorized size classes according to body length and counted individuals of each size category in 10% of the sample. Most zooplankton including small cladocerans and copepods are known to graze efficiently on algae smaller than 30 µm in size (Lampert and Sommer 2007). Therefore, we estimated the fraction of phytoplankton smaller than 30 µm for the major axis of the cell or colony in the total phytoplankton biomass as edible fraction (\(\alpha_{\text{edi}}\)). Carbon biomass of phytoplankton was estimated using the cell abundance and biovolume (μm3) and that of zooplankton using the individual abundance and length-weight relationships. Details of these methods were described in supplemental methods.