The art of optimization of streamflow and precipitation gaging networks is far advanced but little used. This was one conclusion from the Chapman Conference on the Design of Hydrologic Data Networks held in Tucson Arizona, December 11–14, 1978. The environmental movement, with its expanded demand for water quality data has had little effect on research, nor the research on it. All agreed that this was a historical artifact, and that the discussions at the Conference will help redirect future research. Not only should quality rather than quantity of water be studied, but monitoring as well as research networks, ground water as well as surface water. An interesting idea was that if the data network is optimized separately from the data user, and if the data user is using a suboptimal or non-optimal decision rule, more data may produce worse decisions, resulting in a ‘negative worth of data.’ A final major conclusion was that optimization and Bayesian analysis are not sufficient to derive a usable solution. The political framework is as important as the hydrologic framework for decision making. A decision mechanism which is understood and used is better than one which is for mally correct but too complex either to be implemented or to be used.