FISHERY BULLETIN: VOL. 74, NO. 4 



western Oregon. They consisted of two large 

 wooden channels interconnected at the ends by 

 pipes (Figure 1). Within each stream were four 

 riffle-pool sections of equal size; the total surface 

 area available to fish and other organisms was 22 

 m^'. Minor differences in substrate composition, 

 water velocity and depth, and shading from ter- 

 restrial vegetation existed among the riffle-pool 

 sections. These variations were sufficient to 

 prevent the sections from being treated as rep- 

 licates; therefore, samples from each of the four 

 sections were composited. 



The slope of each stream was approximately 

 1.9% so that water pumped into the upstream ends 

 flowed downstream at velocities typical of natural 

 streams (approximately 60 cm/s in the riffles to 

 near cm/s at the bottom of the pools). A 2- 

 horsepower centrifugal pump forced water from 

 the downstream end of one channel to the up- 

 stream end of the other. A gate valve controlled 

 the flow rate, which was maintained at approx- 

 imately 1.35 m'^/min. 



Complete freedom of movement for the fish was 

 allowed between the two channels. Individuals 

 could pass downstream or upstream through the 

 pipe from one side to the other; they were, how- 

 ever, prevented from entering the pump by a 

 screen at the downstream end of the lowermost 

 pool. Movement of the fish from the streams took 

 place through a 6-cm diameter outlet pipe that 

 originated at the screen and terminated in a 

 partitioned trap. Fish that entered the trap were 

 returned to the uppermost riffle both to avoid 

 fortuitous losses and to provide the fish with an 

 adequate opportunity to establish residence. 



Substrate consisted of a layer of rocks approx- 



imately 7 cm deep. Following Cummins' (1962) 

 terminology, cobbles and pebbles composed more 

 than 95% of the substrate, both in the riffles and 

 pools, while larger sand was almost absent. No 

 large boulders were present, although a few cob- 

 bles projected above the water. A difference in the 

 amount of very fine sediments existed between 

 the two streams; this difference will be discussed in 

 connection with their invertebrate faunas. 



Temperature Regulation 



Water temperature in the unheated control 

 followed natural diel and seasonal cycles (Figure 

 2). Two 6-kw stainless steel heaters regulated by a 

 variable input timer facilitated temperature 

 elevation in the heated stream. Continuous 

 recordings of the temperature were made by 

 Partlow RFT thermographs.^ Differences between 

 monthly means ranged from 3.3°C (August 1972) 

 to 4.9°C (December 1972): the average tempera- 

 ture difference between the streams was 3.9°C. 



Both streams received 10-20 liters/min of 

 unfiltered water from a small spring-fed creek 

 that contained aquatic invertebrates and algae, 

 but no fishes. During periods of low stream flow, 

 the water supply was supplemented by a mixture 

 of well water and unfiltered water pumped from a 

 large nearby creek. The model streams have been 

 operating continuously at approximately the same 

 temperature differential since completion of con- 

 struction in 1969 (Iverson 1972). However, in 

 December 1972 unusually cold weather caused 



^Reference to trade names does not imply endorsement by the 

 National Marine Fisheries Service, NOAA. 



Figure L-Top. Plan view of model 

 streams. Arrows indicate direction of 

 water flow and black squares in the 

 heated stream denote location of 

 heating units. Bottom. Cross section 

 of one of the channels in a model 

 stream. 



Eichance 



yi^-si/i^^iSi!^ 



Outflow Exchange 



pcaag;-T^::iE'.C»c> v - 





u 



V^^gigS I , , , , ; 



mmM^^ 



■vti«f''iic-,-;iaB., 



764 



