Results/Conclusions   While the abundance of Two-Lined Salamanders is similar between urban and reference streams after hatch-out in the spring, we have documented consistent decreases in abundance of the species in urban streams as the larval period progresses. We have also recorded lower substrate diversity and lower stream bed cover availability in these same urban stream habitats. To evaluate the hypothesis that decreased stream bed cover causes salamanders to drift downstream, resulting in decreased larval abundance after hatch, we conducted an experiment. We created a flow-through trough that we lined with one of four substrate types found in Piedmont streams (sand, sand/detritus, gravel/pebble mix, or pebble/cobble mix). We predicted that salamanders would be able to withstand higher water velocity on substrates more similar to the reference condition (i.e, high pebble/cobble composition). After an acclimation period, larval salamanders were subjected to periodically increasing water velocity on each of the substrate types. Salamanders were flushed from the flumes with sand and sand/detritus substrates at significantly lower velocities than salamanders on rocky substrates. This experiment demonstrates the importance of in-stream sediment composition on salamander drift. While there are undoubtedly many factors that contribute to salamander mortality in urban habitats, our results suggest that the absence of refugia from high stream flows is a necessary addition to other factors that negatively impact salamander larvae such as water toxicity or loss of adult habitat. Fortunately, the addition of pebble- and cobble-sized substrate particles to stabilize streambeds is already part of many stream restoration efforts. Our data suggest this practice has biotic, as well as abiotic, benefits.