Environmental DNA (eDNA) is used to study the presence, and even abundance, of a range of aquatic organisms. However, our understanding of eDNA transport dynamics in streams and rivers remains limited. Moreover, questions remain about how eDNA characteristics vary among species and how this influences eDNA transport in flowing waters. Here, we build on our previous work in recirculating mesocosms, where pulsed eDNA additions from Common Carp (Cyprinus carpio) and Steelhead Trout (Oncorhynchus mykiss) showed that eDNA particle size distribution and water column removal rates differed between the species; Steelhead eDNA had more small particles and was removed more rapidly across a range of substrate, biofilm, and light treatments. To extend this work to more realistic systems, we conducted short-term, steady-state additions of Carp and Steelhead eDNA in experimental streams to compare the effects of substrate and biofilm on eDNA transport for these species. The 50-m long x 0.4-m wide streams (N=4) were lined with differing substrates, and we compared shaded conditions (90% shade cloth) to open-canopy conditions with distinct biofilm phases (early vs. late colonization). Finally, we partitioned eDNA size classes using sequential filtration through 1.2 and 0.4-μm filters. For each short-term addition (N=32 total), we estimated eDNA removal based on declines in water column eDNA concentrations. Thus far, we found that transport length (Sw) was significantly longer for Carp (Sw=77m) compared to Steelhead (Sw=67m) for early biofilm growth on pea gravel. When converted to depositional velocity (Vdep), Carp and Steelhead removal (Vdep=0.032 and 0.037 mm/sec, respectively) were low compared to mesocosms (Vdep=0.29-2.4mm/s), likely because the experimental streams had lower velocities, were shallower, and did not include biological decay (over 48hr) in addition to physical removal. This work will help to identify controls on eDNA transport, which will advance eDNA as a monitoring tool for aquatic species of concern.