The toxic dinoflagellate Alexandrium fundyense Balech was grown under temperature‐ and nutrient‐limited conditions, and changes in labeling intensity on intact cells were determined for two probe types: an oligonucleotide probe targeting rRNA and a monoclonal antibody (MAb) targeting a cell surface protein. In nutrient‐replete batch culture, labeling with the rRNA probe was up to 400% brighter during exponential phase than during stationary phase, whereas MAb labeling did not change significantly with growth stage at the optimal growth temperature. In cultures grown at suboptimal, low temperatures, there was a significant difference between labeling intensity in stationary versus exponential phase for both probe types, with exponential cells labeling brighter with the rRNA probe and slightly weaker with the MAb. The decrease in rRNA probe labeling with increasing culture age was likely due to lower abundance of the target nucleic acid, as extracted RNA varied in a similar manner. With the MAb and the rRNA probes, slower growing cultures at low, nonoptimal temperature labeled 35% and 50% brighter than cells growing faster at warmer temperatures. Some differences in labeling intensity per cell disappeared when the data were normalized to surface area or volume, which indicated that the number of target antigens or rRNA molecules was relatively constant per unit area or volume, respectively. Slow growth accompanying phosphorus and nitrogen limitation resulted in up to a 400% decrease in labeling intensity with the rRNA probe compared to nutrient‐replete levels, whereas the MAb labeling intensity increased by a maximum of 60%. With both probes, labeling was more intense under phosphorus limitation than under nitrogen limitation, and for all conditions tested, labeling intensity was from 600% to 3600% brighter with the MAb than with the rRNA probe. Thus, it is clear that significant levels of variability in labeling intensity can be expected with both probe types because of the influence of environmental conditions and growth stage on cellular biochemistry, cell size,rRNA levels, and the number or accessibility of cell surface proteins. Of the two probes tested, the rRNA probe was the most variable, suggesting that in automated, whole‐cell assays, it can be used only in a semiquantitative manner. For manual counts, the human eye will likely accommodate the labeling differences. The MAb probe was less variable, and thus should be amenable to both manual and automated counts.