The ''transmission-line'' model of return-stroke radiation, proposed by Uman and McLain (1970) and invoked frequently thereafter to deduce peak currents from remote fields or to estimate propagation velocities from measured fields and currents, has never received a thorough experimental test. During the summer of 1985 at the Kennedy Space Center in Florida, we were able to measure peak currents (with a coaxial shunt), two-dimensional average propagation speeds (with a high-speed streak camera), and electric field waveforms (at 5.15-km range) for a number of subsequent return strokes in rocket-triggered lightning flashes. Because of the temporal ambiguity on the streak-camera films, it has not been possible to identify individual velocity measurements with particular strokes for which current and field data are available. Three multistroke flashes, however, each yielded a tight cluster of velocity measurements and a group of peak field to peak current ratios, though not necessarily for the same strokes. A futher six flashes provided more current and field measurements for which no velocity information was obtained, and velocity measurements only are available for still other flashes. It is shown that these data indicate reasonable agreement between the propagation speeds measured with the streak camera and those deduced from the transmission-line model. The previously observed difference between current and radiation-field waveforms suggests a modification of the model, involving two wave fronts traveling upward and downward away from a junction point a short distance above the ground, which substantially improves the agreement between measured and inferred propagation speeds. ¿ American Geophysical Union 1988