At the initiation of the experiment the amount of fresh substrate to be delivered per unit weight of alkali fed (a) was set on the computer. In the initial series of experiments, a was varied to determine its effect on the dilution rate and residual glucose concentration at steady state. The software also incorporated a short program for signal conditioning or smoothing, using a moving average calculation method, to filter out noise generated in the electrical circuits contained in the
digital weighting balances. Continuous plot of the weight of alkali and fresh substrate reservoirs were displayed on the computer screen during the course of the experiment.

Analytical Methods: Ethanol concentration was estimated by gas chromatography (GC-8 APE; Shimadzu Corporation, Kyoto Japan) with PEG 80-100 mesh. The temperature of the column oven and the injector were 70°C and 90°C, respectively. Glucose concentration was determined with a glucose analyzer (Model 23 A, Yellow Spring Instrument Co. Ltd., Ohio USA). Biomass was estimated using correlation between optical density measurements at 562 nm using a spectrophotometer (Uvidec 320 AS Co. Ltd., Tokyo Japan) and dry cell weights.