From a data base consisting of (1) correlated optical and electric field measurement for 11 natural lightning strokes in Florida, (2) correlated optical and current measurements for 32 artificially initiated (triggered) lightning strokes in New Mexico, and (3) correlated optical and current measurements for 36 triggered lightning strokes in Florida, dart leader speed is examined as a function of the following return stroke initial electric field peak, of the following return stroke current peak, and of the duration of the previous interstroke interval (excluding the duration of continuing current, if present). Return stroke current peaks in both New Mexico and Florida triggered lightning were converted to electric field peaks via the field-current regression equation obtained by Willett et al. (1989), while the electric field peaks in Florida natural lightning were converted to current peaks using the current-field regression equation derived by Rakov et al. (1992), both formulas being based on the same Florida triggered lightning measurements. For each of the three data sets, dart leader speed and the following return stroke field peak or current peak are positively correlated. The relations between leader speed and field or current peak for Florida triggered and Florida natural lightning are similar, possibly indicating a similarity between at least some features of dart leaders and return strokes in natural and in triggered lightning at the same geographic location. On the other hand, leaders at New Mexico triggered lightning are, for the same value of return stroke field or current peak, about twice as fast as those in both triggered and natural lightning in Florida, the difference being likely associated with the relatively short preceding interstroke intervals in New Mexico triggered lightning. For all triggered and natural lightning data taken together, there is a weak but statistically significant tendency for lower leader speed to be associated with a longer previous interstroke interval. However, neither the New Mexico triggered lightning data nor the Florida triggered lightning data, when taken separately, show this tendency. ¿ American Geophysical Union 1992