An eight-input translog total cost function was estimated, and the estimates were used to calculate Allen and Morishima substitution elasticities, price elasticities of conditional input demands, elasticity of size, and technical change for Norwegian sawmilling at a five-digit disaggregate industry level, based on an unbalanced panel of 3,090 individual observations of sawmills over the period 1974–1991. The unusually rich source of data made it possible to analyze the production economics of sawmilling in more detail than has ever been possible before. As far as comparisons with other studies could be made, many of the Allen substitution elasticities and price elasticities of conditional input demand were in accordance with previous findings. The inputs were found to be inseparable. Therefore, frequently used overall aggregates like “materials” and “energy” are not to be recommended. Since the average sawmill had economies of size yet to be exploited, it would have been better off by increasing its size. The most effective way of doing this is to use relatively more sawlogs, electricity, capital, and the aggregate other inputs, and relatively less labor, lumber input (for planing), and the aggregate input other materials. The rate of technical change was negative for the average sawmill. FOR. SCI. 46(4):537–547.