AbstractSediment transport rates in steep mountain channels are typically an order of magnitude lower than predicted by models developed for lowland rivers. One hypothesis for this observation is that particles are more stable in mountain channels due to particle‐particle interlocking or bridging across the channel width. This hypothesis has yet to be tested, however, because we lack direct measurements of particle friction angles in steep mountain channels. Here we address this data gap by directly measuring the minimum force required to dislodge sediment (pebbles to boulders) and the sediment weight in mountain channels using a handheld force gauge. At eight sites in California, with reach‐averaged bed angles ranging from 0.5° to 23° and channel widths ranging from 2 m to 16 m, we show that friction angles in natural streams average 68° and are 16° larger than those typically measured in laboratory experiments, which is likely due to particle interlocking and burial. Results also show that larger grains are disproportionately more stable than predicted by existing models and that grains organized into steps are twice as stable as grains outside of steps. However, the mean particle friction angle does not vary systematically with bed slope. These results do not support systematic increases in friction angle in steeper and narrower channels to explain the observed low sediment transport rates in mountain channels. Instead, the spatial pattern and grain‐size dependence of particle friction angles may indirectly lower transport rates in steep, narrow channels by stabilizing large clasts and channel‐spanning steps, which act as momentum sinks due to form drag.