Riparian cottonwoods (populus spp.) are keystone pioneer species that contribute to critical streamside and in-stream habitats, water quality, and aesthetic and recreational value. Land use and river regulation have caused a widespread reduction in the extent and regeneration of this genus. The majority of research on Populus species' reproduction has examined seedling recruitment that dominates in wide alluvial valleys. In contrast, I evaluated reproductive strategies of Populus angustifoliain mountain valleys. Research was conducted in northern Utah on the Little Bear River, a gravel-bedded stream that flows north out of the Bear River Range onto deposits of Ancient Lake Bonneville. I used allozyme electrophoretic data and vegetation mapping to investigate recruitment mechanisms of cottonwoods at two steep, confined mountain valley sites and two wide, alluvial valley sites. Allozyme electrophoretic analysis on samples from sites one through four revealed 60%, 69%, 86%, and 84% clonal recruitment, respectively. The size of cottonwood clones may be on the order of 200-300 m in this system. Vegetation mapping supported allozyme data and suggested that clonal recruitment dominates the system. I also collected data at four sites containing no cottonwoods to compare hydrogeomorphic influences at non-cottonwood (NC) sites and cottonwood recruitment (CR) sites. CR sites possess a distinctive geomorphic template relative to NC sites and this template may facilitate root disturbance resulting in clonal recruitment. A high frequency of woody debris and mid-channel islands characterizes CR sites. Channels at CR sites are less entrenched, more sinuous, and exhibit larger bankfull channel width and width to depth ratios, and finer grain sizes than channels at NC sites. At CR sites, estimates of bankfull average boundary shear stress and unit stream power are less compared to those at NC sites. Cottonwood recruitment models that typically describe seedling reproduction in alluvial valley environments do not apply in narrow, steep-gradient environments. I found clonal recruitment of cottonwoods to correlate spatially and temporally with channel and root disturbance associated with rain on snow events on the order of the 25-30-yr flood event.