1.8-2.8% (Fairman and Miller 1990). Stream gradients at most Harlequin observation sites in 

 Oregon were between 1-7%. (62% of observations), while most others were in low (<1%) 

 gradient areas (33% of observations) (Thompson et al. 1993). On the Olympic Peninsula of 

 Washington, Harlequins select stream reaches with 1-7% gradient; birds were not present in lower 

 portions or steep headwater sections of occupied streams (Schirato and Sharpe 1992). In Grand 

 Teton National Park, Wyoming, Harlequins used the lowest gradient streams in the park, 

 averaging less than 7% (Wallen 1 987). 



On the Rocky Mountain Front of Montana, Harlequins were observed at sites with stream 

 velocities ranging from 0.8 m/sec to 4.1 m/sec; the average was 1.3 m/sec (n=42) (Diamond and 

 Finnegan 1993). In Idaho, Harlequin Ducks were strongly associated with swiftly flowing 

 streams, however, aduhs were found in higher average velocity (1.2 m/sec) locations than were 

 broods (0.9 m/sec) (Cassirer and Groves 1994). Stream velocity of breeding streams in Iceland 

 ranged from 0.5-3.0 m/sec (Bengtson 1972). 



Channel morphology On McDonald Creek in Glacier National Park, Montana, Harlequins 

 used straight, curved, meandering, and braided stream reaches in proportion to their availability 

 (Ashley 1994). On the Rocky Mountain Front of Montana, most Harlequins used streams that 

 were controlled by v-shaped valleys, either straight (34%) or curved (29%) (Diamond and 

 Finnegan 1993). In Oregon, 45% of sightings were in straight channels (Thompson et al. 1993). 

 In Grand Teton National Park, Wyoming, more than 50% of all Harlequin observations were 

 made in meandering stream channel types (Wallen 1987). 



Aquatic habitat In Glacier National Park, Kuchel (1977) found that Harlequin aduhs 

 selected stream habitats over backwater habitats during the first week of May, but that selection 

 was reversed dunng 12 June - 2 July (p < 0.01). Bottom types were randomly selected except 

 during the period 8-14 May when cobble was selected (Kuchel 1977, Ashley 1994). The 

 strongest selection was for stream reaches with 3+ loafing sites per 10 m (Kuchel 1977, Ashley 

 1994). However, during brood rearing, 83% of observations during the second week occurred in 

 backwaters, which increased to 91% by the time broods were 6 weeks old (Kuchel 1977); this is 

 not evident in Ashley's (1994) data. Broods generally moved downstream as they got older 

 (Kuchel 1977). 



On the Rocky Mountain Front of Montana, most Harlequins used riffles (25%), rapids (24%), 

 or runs (24%), observations were not separated by season (Diamond and Finnegan 1993). 

 Similarly, in Idaho, adults used riffle, run, and rapid areas 74% of the time, while broods used 

 pocketwater and pool/backwater 50% of the time (Cassirer and Groves 1994). In Oregon, adults 

 also used riffle (21%), run (22%), and rapid (22%) areas equally, while pocketwater, pools, 

 glides, and backwaters were used substantially less (Thompson et al. 1993). 



In Iceland, Harlequins overall selected "calm" water as opposed to "fast" or white" water 

 (X^ = 42.4, p < 0.001); however, they preferred to feed in "fast" water (X^ = 32.4, p < 0.001) 

 (Inglisetal. 1989). 



In Montana, bays at the mouths of some breeding streams and lakes are occasionally used by 

 foraging and loafing Harlequins and may provide night roosting areas late in the brood rearing 

 season (Reichel and Genter 1996). There are muhiple records of broods using lakes in British 

 Columbia, but the extent and timing of use is unknown (Campbell et al. 1990, Breault and Savard 

 1991). In Alaska, estuaries and intertidal dehas provided habitat for foraging and loafing 



