Maynard et al.: Movement and mortality of Atlantic salmon kelts (Salmo salar) released into the Penobscot River, Maine 
287 
Representative paths of Atlantic salmon (Salmo salar ) kelts tracked with 
radio telemetry from mid-November 2015 to min-June 2016 on the Penob¬ 
scot River, Maine. The tracks in the top 2 graphs indicate movement of 2 
kelts (from the release site at river kilometer [rkm] 43.5) after the first 
month of release and the tracks in the bottom 2 graphs indicate move¬ 
ment of 2 kelts from the release site at rkm 92. The horizontal dotted 
lines represent dams located at rkm 61 and rkm 99 on the main stem of 
the river. 
Research from European rivers 
suggests that females are generally 
in better condition than males after 
spawning, likely because of low in¬ 
stances of aggression among female 
fish (Halttunen et al., 2013). Because 
their condition is generally better af¬ 
ter spawning, it may be more advan¬ 
tageous for females to remain in the 
low-productivity, low-risk freshwa¬ 
ter environment through the winter 
(Halttunen et al., 2013) to keep the 
risk of mortality low for those months, 
whereas males must enter the marine 
environment sooner to recondition. In 
our study, both males were presumed 
dead. Previous research has also doc¬ 
umented low postspawning survival 
among males, which leads to a higher 
proportion of iteroparous fish being fe¬ 
male (Niemela et al., 2006; Halttunen 
et al., 2010). However, given that the 
sex-ratio of our study fish was so heav¬ 
ily skewed toward females, it is pos¬ 
sible that our results are applicable 
only to female fish. With only 2 male 
fish (both released at the upriver re¬ 
lease, and both mortalities), it is very 
likely that we were unable to capture 
the full range of male postspawning 
behavior. 
Unlike the fish in several other 
studies of kelt movements (Bagliniere 
and Porcher, 1994; Komandina-Douth- 
wright et al., 1997; Cunjak et al., 
1998), many fish in our study made 
directed, upstream movements after release. However, 
in our study, kelts were released downriver from where 
they would naturally spawn, if not taken to the hatch¬ 
ery. This provides at least 2 possible reasons why fish 
would move upstream: to seek more advantageous hab¬ 
itat conditions and because they may have been more 
physiologically similar to ripe adults than postspawn¬ 
ing adults owing to the artificial spawning process at 
the hatchery. 
Further study is needed to clarify the causes of 
upstream movements, such as those we observed in 
hatchery-spawned kelts. One possible explanation for 
upstream movement centers on habitat availability. 
Estuarine habitat may be most desirable for kelts dur¬ 
ing severe winters (Cunjak et al., 1998) because of the 
absence of frazil ice in that area. However, the win¬ 
ter of 2015-16 was unusually mild, which may have 
contributed to kelts overwintering farther upstream 
in the watershed during our study. Additionally, stud¬ 
ies in other systems (Saunders, 1960; Komandina- 
Douthwright et al., 1997; Cunjak et al., 1998) have 
found that habitat complexity (e.g. lakes, large pools, 
islands, and other structure) was positively correlated 
with higher instream survival of postspawning fish and 
higher usage of freshwater habitat instead of marine 
or estuarine habitat. The first 38 rkm of the Penobscot 
River are commercially navigable, and relatively few 
islands (natural or artificial) exist downstream of rkm 
48. In contrast, the reach between rkm 48 and rkm 
61 is filled with artificial islands built for log drives, 
as well as several large natural islands. Therefore, for 
fish released at the downriver site, most of the com¬ 
plex habitat exists upriver of their release point. How¬ 
ever, because the habitat between rkm 61 and rkm 99 
is mostly complex, braided channel with many small 
islands, upstream migration would not be necessary for 
fish released at the upriver location to reach suitable 
overwintering habitat. 
Another possible explanation for upstream migra¬ 
tion after release is physiological. In our study, most 
of the fish in the lower release group and nearly half 
of the fish in the upper release group made directed 
movements upstream after release. However, other 
researchers have found that kelts move downstream 
after spawning (Bagliniere and Porcher, 1994). Up¬ 
stream movements late in the spawning season, as 
documented here, have only been documented in fish 
that were still ripe (Komandina-Douthwright et al., 
