3 males were alive in 1995; of 42 juveniles marked in 1993, at least 1 female and 1 male were 

 alive in 1995. All males known to be alive were seen on the wintering grounds only. 



In Glacier National Park, all mortality of ducklings (through fledging) took place in the 

 first three weeks of life (Kuchel 1977). This is similar to the findings of Bengtson (1966, 1972), 

 who reported that of 7 broods totaling 37 ducklings, 24 survived one week, and 19 survived two 

 weeks; little mortality was seen after two weeks. Bengtson (1972) reported that survival of 

 ducklings ranged from 40-76% on 3 streams over 5 years. An extreme example showing little 

 mortality after 1-2 weeks is a brood of 5 Class lb young (8-15 days-old), which was first seen on 

 Marten Creek, Montana, on 10 July 1995, without an adult female present. All survived and 

 were nearly flying on 28 July 1995. 



In Idaho, 63% of adults (n=30) returned at least 1 year; male and female rates were not 

 significantly different (Cassirer and Groves 1994); one duck marked as an adult in 1988 returned 

 through 1993 (minimum 7 years old). No ducklings marked from 1988-1991 were re-observed 

 (n=27). In Wyoming, 40% of marked adults returned to breeding streams (Wallen 1993). At 

 least 5 females of 103 ducklings banded in 1987-1990 have returned and nested successfully 

 (Wallen 1991). The oldest known Wyoming bird was marked as a duckling in 1985 and 

 recaptured in 1991 (Wallen 1993). In Alaska, 30% (8) of adult females and 30% (3) of adult 

 males marked were relocated the following year (Dzinbal 1982:62). 



In Iceland, 64%i (20) of adult females and 48%) (13) of adult males, marked with nasal 

 discs, were relocated the following year (Bengtson 1972). Hatching success in Iceland averaged 

 87%, and ranged from 84% to 91% in four years (Bengtson 1972). 



CAUSES OF MORTALITY 



Causes of death. Reproduction in Montana in 1995 was one of the poorest on record 

 (Table 5, 6). Table 5 shows the North Fork Flathead drainage reproduction to be almost non- 

 existent, while reproduction on Lower Clark Fork streams was near average. An extremely 

 heavy spring storm in the North Fork caused streams to exceed 1 00-year flood levels (M. White 

 pers. comm. in Ashley 1995); this heavy storm and flood did not hit the Lower Clark Fork area. 

 Throughout Montana, high water during early summer runoff has been associated with low 

 productivity by a number of studies (Kuchel 1977; Diamond and Finnegan 1992, 1993; Reichel 

 and Center 1993, 1995). Possible explanations for the negative correlation of survival with run- 

 off include females not nesting due to high water and/or poor feeding; destruction of nests within 

 the floodplain; or loss of juveniles due to drowning, separation fi-om the female, inability to feed 

 effectively, or hypothermia. 



In Idaho, productivity was negatively correlated to June stream flow (r = -0.93, p = 0.006) 

 (Cassirer and Groves 1995). Wallen (1987) reported that neither of two broods seen prior to a 

 severe July rainstorm, which raised a creek level 0.6 m within 2 hours, was ever seen again; 

 however, he generally felt that drought in the Grand Tetons was more limiting to reproductive 

 success than was flooding. Dzinbal (1982) reported that higher spring run-off was associated 

 with lower reproduction in a two-year study in Alaska. 



Bengtson (1972) found that very low duckling survival coincided with adverse weather 

 and with very low abundance of blackflies, the preferred food in the study area. 



In coastal waters. Harlequins are occasionally caught by the bill and drovmed by large 

 mussels and clams (Turner 1886 in Philips 1925). 



19 



