AbstractIn April—May, 1971 and 1972, temperatures measured in Quito, Ecuador, at a height of 1.4 m were characterised by horizontal differences which contrasted strongly with the generally accepted patterns for mid-latitude cities. The separated heat islands over the old and modern sections of the city achieved their greatest magnitudes between 1200 and 1300 hours solar time, and these were frequently 5–10 times greater than the mid-day values reported from mid-latitude locations. On the other hand, in the three-hour period preceeding midnight, when heat islands in mid-latitude cities generally achieve their greatest magnitudes, those in Quito were obliterated by the drainage of cooler air from the very steep slopes to the west. By day, the horizontal temperature gradients between Quito's parks and open spaces on the one hand and its modern section on the other were much steeper than any which have been previously reported from other locations. A previous analysis of those critical wind velocities which, in seven mid-latitude cities, caused the destruction of heat islands, indicated that, if no additional factor was involved, such features in Quito should have been obliterated when the wind speed reached 7.9 m/sec. However, there were several occasions on which winds of 9–11 m/sec. permitted well-developed heat islands to remain as dominant features of the daytime temperature pattern. An examination of the possible reasons for the differences in magnitude between the two separated parts of Quito's heat island indicated that albedo was the dominant factor. Finally, it is suggested that the results of an urban climatic study should only be extrapolated within the same macroclimatic region and to areas where the local topography is essentially similar.