This study is concerned with the estimation of economies of scale in thermal-electric power systems subjected to water quality standards. The approach is engineering-economic. Mixed-integer programming is used to solve a mathematical model of a power system subject to constraints defining (1) the load duration curves of several demand centers located on a spatial grid or network, (2) endogenously determined reserves equal to the largest two generating units in the system, (3) nonlinear transmission losses, and (4) environmental quality and emission standards. The objective of the model is to minimize the sum of total capital and operating costs for generation, transmission and abatement of pollution emissions. Fixed charges and piecewise linear functions are used to define costs exhibiting increasing returns to scale. The model is applied to a major part of the New York State Electric and Gas Corporation system. (Modified author abstract)