Plant cell aggregates have long been implicated in affecting cellular metabolism in suspension culture, yet the rigorous characterization of aggregate size as a process variable and its effect on bioprocess performance has not been demonstrated. Aggregate fractionation and analysis of biomass-associated product is commonly used to assess the effect of aggregation, but we establish that this method is flawed under certain conditions and does not necessarily agree with comprehensive studies of total culture performance. Leveraging recent advances to routinely measure aggregate size distributions, we developed a simple method to manipulate aggregate size and evaluate its effect on the culture as a whole, and found that Taxus suspension cultures with smaller aggregates produced significantly more paclitaxel than cultures with larger aggregates in two cell lines over a range of aggregate sizes, and where biomass accumulation was equivalent before elicitation with methyl jasmonate. Taxus cuspidata (T. cuspidata) P93AF cultures with mean aggregate sizes of 690 and 1,100 mu m produced 22 and 11 mg/L paclitaxel, respectively, a twofold increase for smaller aggregates, and T. cuspidata P991 cultures with mean aggregate sizes of 400 and 840 mu m produced 6 and 0.3 mg/L paclitaxel, respectively, an increase of 20-fold for smaller aggregates. These results demonstrate the importance of validating experiments aimed at a specific phenomenon with total process studies, and provide a basis for treating aggregate size as a targeted process variable for rational control strategies. (C) 2011 American Institute of Chemical Engineers Biotechnol. Prog., 27: 1365-1372, 2011