The experiment was conducted at the University of Michigan Biological Station Artificial Stream Facility on the East Branch of the Maple River in Emmett County, Michigan. Mayflies and stoneflies were collected from various locations on the river using dip nets and kick nets by walking backward upstream and disturbing the substrate. They were then randomly sorted into groups of ten (mayflies) and two (stoneflies). Channel construction was based on the experiment set-up from Joe Holomuzki’s (1996) experiment conducted at the UMBS Stream Facility. Vinyl rain gutters (2.8 m long, 0.1 m wide, 7 cm high) were used as artificial stream channels. Water was pumped from the Maple River by a Monarch® pump with 2.54 cm holes into a plastic barrel with 8 valves (“head tank”) that each emptied into a channel and regulated flow. Water in each channel flowed through a “pool” created by a dam constructed of four clay tiles (4 x 3 cm) stacked 2 tiles high that collected incoming FPOM (free particulate organic matter), then through a meter-long section containing substrate and mayflies 0.3 m from the substrate. Drifting mayflies were captured by a 5.08-cm section of PVC pipe with a piece of window screen glued to one end that was attached to each gutter with a short section of K-mart® brand knee-high nylons. Discharge (120 mL/s, with an additional 60 mL/s for the increased flow treatments, based on average flow fluctuations collected by USGS data for nearby streams) was uniform among all streams. pH and temperature were measured in the river and in each channel before each trial. Mayflies were poured gently into each channel from a paper cup behind a handheld clay tile (5 x 5 cm) that served as a dam to allow the mayflies to settle and was removed once all ten mayflies settled. They were added at sunset and collected and counted the next morning, then sorted based on treatment and whether they drifted and preserved in 95% ethanol. For the four high flow treatments, an extra head tank filled by a sump pump that was connected to a timer was turned on later in the trial. For the treatments containing predators, 2 stoneflies were added to each channel using the same method as mayflies at the same time as the increase in flow. Substrate (sand, gravel, and cobble) was collected from the Maple River and placed in the last meter of each channel. Sand and gravel were spread out to make a smooth layer approximately 1 cm deep, and 6-7 pieces of cobble were placed on top of the gravel in an alternating pattern. Drift rate for each treatment was calculated as the number of mayflies that drifted from each trial divided by the total number that drifted from all treatments. Drift rates were compared between treatments using SPSS to conduct an independent t-test analysis, and the interaction between the different variables was calculated with ANOVA. There will be no chemical manipulations to the water, and no biological ones other than disturbance caused by collecting animal subjects and filtering substrate and organisms from the water that is pumped from the stream.