To test whether total-tract starch digestibility by lactating dairy cattle could be predicted accurately from concentration of starch in fecal dry matter (FS), data from 8 digestibility trials, 30 diets, and 564 individual starch measurements were compiled from trials conducted at the University of Wisconsin-Madison. Total-tract starch digestibility by individual cows was determined from the concentrations of starch in total mixed rations and feces (rectal grab samples) and concentrations of several external or intrinsic feed markers. Fecal starch concentration was closely and linearly related to total-tract starch digestibility (R(2) = 0.94). Differences in fecal sampling time for FS, such as day within week or week, did not influence determination of FS concentration. In contrast, time of day when fecal samples were taken influenced FS concentration, but the differences were minimal (mean ± standard error of the mean = 0.4 ± 0.1). These data suggest that on-farm collection of feces from individual cows or pens of cows may be sampled only once per day. Fecal pH was not related to FS when FS was only 1 to 3%, indicating that fecal pH was not a good index of FS and not practical as a tool to assess total-tract starch digestibility for lactating dairy cows when fecal starch excretion is low. An equation was developed to predict FS using near-infrared reflectance spectroscopy of dried ground fecal samples; the equation had moderate to good accuracy (R(2) = 0.83-0.94) and a low standard error of prediction. Fecal starch concentration can be used to monitor total-tract starch digestibility.