Functional wing length (wing length/head-capsule width) of female Sweltsa revelstoka (Jewett) from streams was measured for 19 sites that have been free of Wisconsin glacial ice for about 15 000 years and possibly longer, and from 23 sites that have been ice free for about 10 000 years. At the former sites brachypterous populations were common and there was a significant negative relationship between functional wing length and elevation, and a positive relationship between functional wing length and stream size. In the area that deglaciated more recently, populations were not or only slightly brachypterous and there was no significant relationship between wing length and elevation or between wing length and stream size. Functional wing length was not related to body size. These analyses indicate that the brachypterous condition is probably genotypic in origin. I suggest that streams were colonized by macropterous forms shortly after deglaciation, and that brachyptery takes several millennia to develop at small, high-elevation streams.