Recent experimental ecosystem studies on discrete watersheds have greatly increased our understanding of ecosystem nutrient cycling. For our purposes a watershed can be defined as a landscape catchment basin, including terrestrial slopes, streams, and lakes, from which all runoff comes out through a common stream outlet. In terms of study and management by man a watershed is a convenient – sized ecosystem with definable boundaries. Especially interesting research is being carried out at the Hubbard brook experimental watersheds in new Hampshire, the Coweeta watersheds in western north Carolina, and similar research areas in other geographical regions of this continent and elsewhere in the world. In these outdoor laboratories the inputs and outputs of water, nutrients, energy, and so on are carefully measured Hubbard brook watersheds that is covered with an undisturbed forest. Retention by, and recycling within, the undisturbed forest is so effective that the estimated loss from the ecosystem is only 8 kg/ hectare/ year of calcium (and equally small amounts of other ecosystem nutrients). Since 3 kg/ ha of this is replaced in rain, only an input of 5 kg/ ha is needed to achieve a balance, and this is supplied by the normal rate of weathering of the underlying rock that constitutes the ‘’ecosystem reservoir pool.’’ When the vegetation on one of the experimental watersheds was felled and regrowth the next season was suppressed by herbicides, stream outflow increased and the loss of nutrients was 3 to 16 times (depending on the nutrient) that of adjacent undisturbed watersheds. This large – scale experiment illustrates a common situation where a manipulation tht changes the movement of one material (water, in this case) along one pathway also results in changes in other material flows. Increased water yield following removal or reduction of vegetation cover might be advantageous to a city downstream that is short of water, but if such an increased water supply comes at the expense of impoverished productivity of the land (due to unbalanced ecosystem nutrient budget) and a decrease in water quality downstream due to increase in mineral and sediment content), then there be no net advantage to man. Accordingly, for ever proposed manipulation we must inquire into the ‘’trade – off’’ costs. This is, in fact, what is now required by the national environmental policy act (NEPA) recently passed by the U.S. Congress.