Developing accurate scaled-down fermentation platforms that are predictive of large-scale performance and simultaneously detect incremental improvements in yield can be challenging for mycelial (viscous) fermentations, especially those utilizing industrial grade medium (low cost, high non-dissolved solids content).  Small scale micro titer plates and shake flasks used for mutant selection are often subject to mass transfer limitations that can mask higher yielding strains or give false positives.  Difficult to scale parameters such as shear rate can have a dramatic impact on cell aggregation, altering viscosity and thus mixing characteristics. Once a predictive model is established, detecting mutants with improved yield is often accomplished in small, incremental steps.  System variance must be known and minimized in order to drive statistical significance of experiments.  This minimizes the number of required replicates and speeds up the tech transfer process to production.  This presentation will describe not only methods that have been employed at Dow AgroSciences to help detect and overcome scale-up limitations but will also examine experimental considerations that can minimize process variability and improve statistical significance of experiments.