The thinking and experience that went into the ORNL-ANP heat exchanger development program for high-performance heat exchangers has brought out many points having broad implications for other types of high-temperature heat exchanger. This paper summarizes the design precepts and philosophy on which this work was based. While it is evident that weight and volume are vitally important in aircraft power plants, there is also a strong incentive in stationary and marine power plants to reduce both weight and volume because of such considerations as shielding, remote handling, liquid inventory, reactor hazards, control response rates, costs, etc. Analysis disclosed that the tube diameter should be as small as possible consistent with limitations imposed by deposits on the tube walls. Test experience demonstrated the practicality of tube diameters from ¼ to ⅛ in. o.d. It was found that thermal stresses imposed the most important single set of fundamental limitations on the heat exchanger design, and that thermal strain cycling associated with changes from low to high power was the most important failure mechanism. This, coupled with leak tightness requirements, made it essential that a ductile material be employed. The metal also had to be well suited to both welding and brazing because the only thoroughly satisfactory tube-to-header joints tested were first welded and then back-brazed. A series of heat exchangers designed according to these precepts was built and endurance tested at power densities as high as 10 Mw/ft3 (350 kw/liter). Many of the units were endurance tested for over 1000 hr at temperatures up to 1500°F.