1994). Likewise, in Glacier National Park, ducks were seen primarily in areas with old-growth or 

 mature overstory (90%); note however that this figure is approximately in proportion with habitat 

 availability (Ashley 1994). Near Prince William Sound, Alaska, 10 nests were located in 1991-93, 

 they were all associated with old-growth forest (Crowley 1994). In Oregon, however, old- 

 growth accounted for only 48% of sightings (Thompson et al. 1993). On the Rocky Mountain 

 Front of Montana, more birds were seen in pole sized timber (38%) than in other size classes, 

 including mature and old-growth combined (26%) (Diamond and Finnegan 1993); tree 

 composition was evenly split between Douglas fir, lodgepole pine and Englemann spruce. On the 

 Olympic Peninsula of Washington, there appeared to be no selection of particular forest types 

 (Schirato and Sharpe 1992). 



Stream reaches where Harlequins were observed in Idaho were usually not logged or had an 

 unlogged buffer along the stream (Cassirer and Groves 1994). 



Bank composition. In Glacier National Park, most Harlequin observations were adjacent to 

 banks covered with a combination of trees and shrubs; because this was also the most common 

 bank cover type in the park, selection was not evident (Ashley 1994). However, pairs did select 

 for areas with overhanging vegetation (A"^ = 5. 185, p = 0.023) and against areas with undercut 

 banks {X^ = 4.596, p=0.032) (Ashley 1994). On the Rocky Mountain Front of Montana, birds 

 were most commonly seen along banks covered with shrub/tree mosaic (23%) or gravel (20%) 

 (Diamond and Finnegan 1993); locations for most observations did not have overhanging 

 vegetation (70%) or undercut banks (81%). In Idaho, aduh and brood use of banks, by cover 

 type, was similar: trees 40%, shrubs 21%, mosaic 20%, grass/forb 7%, rock/sand/silt 6%, and 

 woody debris 1% (Cassirer and Groves 1994). In Oregon, most banks had trees (41%) or shrubs 

 (25%) present (Thompson et al. 1993). Undercut banks were present at only 18% of sighting 

 locations in Oregon (Thompson et al. 1993). Overhanging vegetation was present at 57% of 

 sighting locations in Idaho (Cassirer and Groves 1994) and 73% in Oregon (Thompson et al. 

 1993). 



In Grand Teton National Park, Harlequins selected banks dominated by shrubs (71%) and 

 sigmcantly avoided banks dominated by grass/forbs (3%) or trees (1 1%) (Wallen 1987:66). 



BREEDING RANGE: AQUATIC 



Stream width. On McDonald Creek in Glacier National Park, Montana, Harlequins appeared 

 to select narrow (<15 m wide) stream locations, although the average stream width was 19 m 

 (Ashley 1994). In western Montana, 7 other streams with Harlequins ranged fi-om 5-30 m in 

 width (Fairman and Miller 1990). On the Rocky Mountain Front of Montana, most Harlequins 

 used streams 6-10 m wide (48%), while the rest used larger (26%) and smaller (25%) streams 

 (Diamond and Finnegan 1993). 



Stream width at most Harlequin observation sites in Idaho was less than 10 m; streams used 

 by juveniles were significantly smaller than those used by aduhs (Cassirer and Groves 1994), 

 though a seasonal adjustment was not made. Stream width at most Harlequin observation sites in 

 Oregon averaged 13.5 m (range 2-50 m) (Thompson et al. 1993). Stream widths of breeding 

 streams in Iceland ranged from 2-40 m; those streams were typically shallow (0.5- 1.0 m) 

 (Bengtson 1972). 



Stream gradient In western Montana, 7 streams with Harlequins had gradients ranging fi-om 



