Work has been carried forward in three different areas. A model of cochlear processing of auditory stimuli is under development. Hair cell transduction is the only nonlinear component in the model and this suffices to explain well known nonlinear auditory characteristics. The second area involves a study of cellular automata. Cellular automata provide a unique approach to complex biological problems such as growth and differentiation of organisms and neural structure and function. Much current interest focuses on general properties or approaches to cellular automata. We have developed an analytical procedure for predicting the statistical distribution of local patterns in a class of simple automatons. The method gives promise of generalization to more complex systems. The third area of study is the statistical properties of biological macromolecules. Relaxed constraints on DNA at amino acid replacements have been discovered. The relative frequencies in the use of synonymous codons in DNA has been analyzed. These frequencies should change under the fixation of mutations unless they are in dynamic balance. Such a dynamic balance has been found for certain classes of sequences and this has led to an estimation of fixation rates in such cases. The PAM matrix model of protein evolution has been analyzed and an error discovered in the construction of the matrix and a need to modify the model demonstrated. An analysis is under way to determine whether the frequency of double amino acid replacements is quadratic in its dependence on divergence time as predicted by the PAM matrix type model of protein change. Finally, a study is progressing to determine the extent of influence of preservation of one residue on the preservation of an adjacent residue, where adjacent may mean close on the linear molecule or close in 3-dimensional space and where preservation is based on the comparison of homologous proteins.
