How many cells should we count?

These data were gathered in an effort to solve a problem with correlating microscopic cell counts and fluorescence measurements taken during the summer of 2000 from an oyster (Crassostrea virginica) grazing experiment reported in Thessen, A.E., Q. Dortch, T.M. Soniat and G.J. Doucette (2009) Crassostrea virginica grazing on toxic and non-toxic diatoms. Toxicon.doi:10.1016/j.toxicon.2009.10.011. The cells in question are the diatom Pseudo-nitzschia. We wanted to answer the question, "Are we counting enough fields and cells to be representative of the sample?" We pulled several slides (slide # 627, 633, 700, 688, 682 and 684) from the experiment of varying cell concentration and species, initially counted at different magnifications. We calculated mean and standard deviation for each field until the two values seemed to stabilize. These data did not solve the correlation problem. We later concluded that a fungal infection in the Pseudo-nitzschia culture had caused the problem.