Local correlations between topography and width behavior within lava flows at Puu Oo, Mount Etna, Glass Mountain, Cerro Bayo, Alba Patera, Tyrrhena Patera, Elysium Mons, and Olympus Mons were investigated. For each flow, width and slope data were both referenced via downflow distance as a sequence of points; the data were then divided into collections of adjacent three-point features and two-point segments. Four discrete types of analyses were conducted: (1) Three-point analysis examined positional correlations between width and slope features, (2) two-point analysis did the same for flow segments, (3) mean slope analysis included segment slope comparisons, and (4) sudden width behavior analysis measured abruptness of width changes. The distribution of types of correlations compared to random combinations of features and segments does not suggest a significant correlation between flow widths and local underlying slopes and indicates that for these flows at least, other factors have more influence on changes in width than changes in underlying topography. Mean slopes underlying narrowing, widening, and constant flow width segments were calculated. An inverse correlation between slope and width was found only at Mount Etna, where slopes underlying narrowing segments were greater than those underlying widening in 62% of the examined flows. For the majority of flows at Mount Etna, Puu Oo, and Olympus Mons, slopes were actually greatest under constant width segments; this may imply a topographically dependent resistance to width changes. The rate of change of width was also examined. Sudden width changes are relatively common at Puu Oo, Mount Etna, Elysium Mons, and Tyrrhena Patera and relatively rare at Glass Mountain, Cerro Bayo, Olympus Mons, and Alba Patera. After correction for mapping scale, Puu Oo, Mount Etna, Olympus Mons, and Alba Patera appear to fall on the same trend; Glass Mount exhibits unusually small amounts of sudden width behavior, and Tyrrhena Patera exhibits a relatively large number of sudden width behavior occurrences. ¿ 2000 American Geophysical Union