The water column was sampled at the deepest point in each lake with a 2.21 Niskin bottle, through a hole drilled in thick ice cover (approximately 4m thick). Duplicate 60ml samples from each depth were fixed in buffered glutaraldehyde to a final concentration of 2% and stored in the dark at 4 degrees C prior to being analyzed. For HNAN and cryptophyte counts, 30-50ml of the sample was stained with DAPI, filtered onto a 2.0 m polycarbonate membrane filter and viewed under epifluorescence. Twenty Whipple grids were counted on each filter to determine mean abundance. For ciliates, duplicate 500 ml water samples were fixed in Lugol's iodine and concentrated by settling, prior to counting in a Sedgewick-Rafter counting chamber under phase microscopy at x160.