A primary goal of ongoing field research of bridge scour is improvement of scour-prediction equations so that pier-scour depth is predicted accurately-an important element of hydraulic analysis and design of highway bridges that cross streams, rivers, and other waterways. Scour depth for piers in streambeds with a mixture of sand, gravel, cobbles, and boulders (coarse-bed streams, which are common in Montana) generally is less than the scour depth in finer-grained (sandy) streambeds under similar conditions. That difference is attributed to an armor layer of coarser material. Pier-scour data from the U.S. Geological Survey were used in this study to develop a bed-material correction factor, which was incorporated into the Federal Highway Administration's recommended equation for computing pier scour. This report describes results of a study of pier scour in coarse-bed streams at 59 bridge sites during 2001-2007 in the mountain and foothill regions of western Montana. Respective drainage areas ranged from about 3 square miles (mi²) to almost 20,000 mi². Data collected and analyzed for this study included 103 pier-scour measurements; the report further describes data collection, shows expansion of the national coarse pier-scour database, discusses use of the new data in evaluation of relative accuracy of various predictive equations, and demonstrates how differences in size and gradation between surface bed material and shallow-subsurface bed material might relate to pier scour. Nearly all measurements were made under clear-water conditions with no incoming sediment supply to the bridge opening. Half of the measurements showed approach velocities that equaled or surpassed the critical velocity for incipient motion of bed material, possibly indicating that measurements were made very near the threshold between clear-water and live-bed scour, where maximum scour was shown in laboratory studies. Data collected in this study were compared to selected pier-scour data from the nationwide Bridge Scour Data Management System (BSDMS), to show the effect of bed-material size and gradation on scour depth. Unsteady field flow conditions and armoring by coarser material reduced scour relative to the clear-water/sandy-bed laboratory results at steady flow. The new correction factor and the standard scour equation produced the most accurate estimates of scour depth in armored, coarse-bed conditions. Maximum relative scour occurred at similar velocity across variations in bed material and gradation. Pier scour decreased with increased variation in particle size and gradation.