The six papers in this “mini-issue” are primarily concerned with the synthesis of digital autopilot logic for adaptive stabilization of straightand-level flight. Some of the logic developed in these papers will be implemented on the unique digital fly-by-wire control system of NASA’s F-8 research aircraft as part of an advanced control law research program sponsored by NASA’s Langley and Dryden Research Centers. The need for adaptive control on aircraft has been debated for many years. The alternative of scheduling autopilot gains accordmg to sensed pressure altitude and Mach number (air data) appears to work well for most aircraft and is relatively simple to implement. To my knowledge the X-15 was the only previous aircraft that had an adaptive control system. While it worked reasonably well, there were problems [I]. Curiously, there is little mention of that project in these papers. Adaptive control appears attractive to handle less predictable things such as changes in atmospheric gustiness @articularly wind shears on final approach) and changes to aircraft weight and mass center (e.g., due to unusual loading distributions or dropping of external stores). To make the present task more challenging, some of the authors assumed only rate gyro and accelerometer information were available. NASA stipulated that only subliminal test inputs could be used for identification. Surprisingly, even with these constraints, key quantities like dynamic pressure and pitching moment due to angle of attack are identifiable to about 10 percent accuracy, which is adequate for the designated task. However, it seems unlikely that a commercial or private pilot would fly without functioning air data sensors, and a military pilot would not be too concerned about detectable test inputs if his air data sensors had failed. Furthermore, an aircraft that could afford adaptive control could probably afford inertial measurement units which give accurate attitude and velocity information. Detection and identification of sensor failures is the subject of one of the papers. The need for this type of adaptation is much less controversial and is part of a larger area of concern, namely, designing for reliability. Detection and identification of engine and actuator failures is another subject of current interest. Developing improved methods for identifying stability derivatives from flight tests of new aircraft is also not too controversial. In this connection it is interesting that the present authors clearly recognize the importance of identifying an “equilibrium offset vector” or “trim disturbance” vector. If control were the objective instead of identification, such constant disturbances are, of course, readily handled using integral-error feedback or lag compensation. I believe it is sigmficant that the authors find it important for identification to use either: 1) “fading memory” or “moving window” algorithms or 2) “fictitious noise” such as Brownian motion models for stability derivatives. This appears to be associated with the “undisturbable mode” difficulty which occurs in the design of steady-state Kalman filters [2]. NASA, IFXE, and the present authors are to be commended for this interesting start on using advanced control concepts in an expenmental aircraft. I hope the program continues and gets into other areas such as sample rate reduction, automatic landing, controlling flexible aircraft, gust alleviation, and controlling aircraft with many inputs and outputs.