The paper describes progress in the development of reliable and practicable online estimations of biomass and secondary product (idiolite) formation in batch processes. Earlier work using data from cellulose and sorbose fermentations is reviewed, describing the success of the application of the Kalman filter to biomass estimation in batch culture. Progress in model development for secondary product formation is also reviewed. Results of investigations into the application of estimation to the batch growth ofStreptomyces clavuligerus(a typical industrial lactam antibiotic producer) are given. A large data base of experimental results for offline model development has been established, and the effects of oxygen-limited growth, stirrer speed and other possible interactions causing deviations from Monod-type growth are discussed. The interactions between process variables are considered, and investigated by corresponding alternative exploratory changes in the model. Implications for control and estimation strategies are explored and preliminary results given for the application of these to the batch process, using prototype hardware and software for online estimation.