During the sixth quarter, the reactor configuration was unchanged but the water-cooled collector is now being cleaned after each run to extract tar (and char) from the collector. This brings the material balance above 90% even for high tar coals. Data for a Beulah, North Dakota lignite were collected in the reactor system at a furnace temperature of 820/sup 0/C. The data suggests that some adjustment must be made to the tar rate or to the particle heating calculation at this temperature. Experiments will be performed at 650/sup 0/C to investigate this problem further. Programmed slow heating-rate experiments were performed for several coals. The use of slow heating in pyrolysis experiments provides data on individual reactions which often cannot be distinguished in rapid pyrolysis because they occur simultaneously. Slow heating conditions often resolve such reactions into separate gas evolution peaks. These peaks provide low temperature data for gas evolution kinetics. Work was performed on the pyrolysis theory. The computational details of the pyrolysis model have been modified in order to allow the simulations of multi-molecular kinetics. This was motivated by the need for more accurate modeling of tar and HCN production. The tar evolution rate is determined by (1) the availability of donatable hydrogen from alipathic or hydroaromatic hydrogen in the char, and (2) by the rate of bridge bond breaking. Similarly, the HCN gas production from the coal depends on both the rate of formation of CN radicals, and the availability of aromatic hydrogen to form HCN. Work continued on elucidating the mechanisms of particle swelling. The effects of gas diffusion through the coal metaplast and of surface tension are considered. The report also considers tar formation which is thought to be the result of a depolymerization process that releases fragments of the coal molecule. 32 figs., 18 tabs.