Sites with drainage areas ranging from 88 to 12,200 km2 were monitored on five streams in northern Alaska during the breakup in 1976 to determine (1) the effects of frozen bed and bank material on channel behavior, and (2) the importance of the annual breakup flood in forming the channels of arctic streams. The thawing and concomitant erosion of channels varied with changes in bed-material size, channel pattern, drainage area, and climate. The response of channels to breakup flooding ranged from total permafrost control of channel processes, including both bed scour and lateral erosion, to only brief restriction of channel behavior early in the rise of the flooding. The watershed characteristic that appears to explain much of this variation is size of drainage area. Similar variation of channel response with these factors is believed to be responsible for diametrically opposite results reported from recent studies of the two problems posed above. Permafrost has been cited both as the cause of extreme stability and as the cause of un usual instability in arctic streams as compared to those elsewhere. That permafrost simplifies the study of arctic stream channels through its domination of the effects of other variables has been a common assumption. As a consequence, however, the generalizations based on a single stream or on similar streams have led to a spectrum of inconsistent results. This spectrum of previous results now ap pears potentially consistent. Comparisons of absolute rates of lateral erosion are not feasible, but it is likely that the net effect of the permafrost environment is to create greater channel stability than is found in unregulated streams of similar size in nonpermafrost environments. Combina tions of factors, particularly those that encourage high rates of thermo-erosional niching, can nevertheless cause rates of erosion that dictate caution in engineering design.