286 
Fishery Bulletin 11 6(3-4) 
3-i 
2 - 
to 
Ll 
c 
1 - 
□ Upriver 
■ Downriver 
1 23456789 
Months since release 
Figure 3 
Timing of outmigration (in months after release) of kelt At¬ 
lantic salmon (Salmo salar ) released in an upriver site (light 
gray, river kilometer [rkm] 92) and a downriver site (dark 
gray, rkm 43.5) on the Penobscot River, Maine, on 17 Novem¬ 
ber 2015 and tracked with radio telemetry from release to 
mid-June 2016. The values on the y-axis represent the num¬ 
ber of fish exiting each month. 
upstream movement. Apparent overwinter survival 
was below 50%; 26 kelts were presumed dead in the 
river system and an additional 16 fish had an unknown 
outcome (8 from each release group). 
The apparent survival for kelts released upriver was 
43%. The maximum distance the 28 upper release fish 
could have moved upstream was 7 km before reaching 
the West Enfield Dam. Further movement was possible 
into the Passadumkeag River for up to 17 km before 
reaching the Pumpkin Hill Dam, although because of 
the absence of roads, this area was not tracked by a 
mobile receiver, except during our one flight. We found 
that 13 fish moved upstream from the release site (11 
during the first month following release), and 10 of 
these approached the dam. Half of the fish (14/28) re¬ 
leased upstream moved downstream past the Milford 
Dam. Most fish that successfully passed the dam did 
so before March, and only one fish descended from the 
upper river in May. Of the fish in this release group, 
4 were documented leaving the system through the 
estuary, including 2 of those that initially moved up¬ 
stream (Fig. 3). Of the 16 presumed fish that died in 
the upriver release group, 6 initially moved upstream, 
8 initially moved downstream, and 2 remained within 
1 rkm of the release site—a finding that indicated no 
relationship between initial movement choice and sur¬ 
vival. Of the presumed mortalities, 6 occurred after 
the fish had moved downstream over the Milford 
Dam. For the purposes of these analyses, mor¬ 
tality and movement are not mutually exclusive 
(i.e. a fish that was counted as moving upstream 
could also be counted as a mortality). 
The maximum distance that the 27 lower re¬ 
lease fish could have moved upstream from the 
release point in Brewer was 17.5 km before 
reaching Milford Dam. Of the 23 fish that moved 
upstream from the release site; 13 of those ap¬ 
proached the dam, and 1 passed the dam during 
a period of extreme high water in mid-December 
2015, when river discharges were more than twice 
the 113-year median. All upstream movements 
originating from the downriver release site took 
place before the end of January. Of these fish 9 
were eventually documented leaving the system, 
including 5 that initially moved upstream (Fig. 
3). An additional 10 fish in this release group 
were presumed dead. Of these fish, 6 moved up¬ 
stream initially, and 3 remained within 1 rkm of 
the release site. 
A larger proportion of kelts released downriver 
were observed making directed upstream move¬ 
ments than the proportion of kelts released at 
the upriver location exhibiting the same behav¬ 
ior (85% vs. 46%, x 2 =7.498, df=l, P-0.006). There 
was no difference in apparent survival to the es¬ 
tuary between the upriver and downriver release 
groups (14% and 33%, respectively, x 2 =l-8084, 
df=l, P=0.179). Apparent mortality over the 
course of the study and apparent initial mortality 
(fish that remained within 1 rkm of the release 
site for the duration of the study) were not different 
for the upriver and downriver groups (Table 2). There 
was no difference in body condition index between the 
fish that were documented successfully leaving the sys¬ 
tem, and the fish that were presumed dead (Wilcoxon- 
Mann-Whitney [W]=367, P- 0.899). There was also no 
difference in body condition index between the fish that 
made directed movements upstream and those that 
moved downstream (W=343, P=0.566). 
Discussion 
The results of this study suggest that the majority (88%) 
of kelts released into the Penobscot River overwinter in 
freshwater, regardless of where they are released. This 
finding is consistent with research from the River Teno, 
where about one third of individuals (35%) moved direct¬ 
ly out to sea (Halttunen et al., 2010). Unlike Halttunen 
et al. (2013), we found no differences in body condi¬ 
tion index between fish that emigrated quickly and fish 
that overwintered in the Penobscot River. However, our 
data set of outmigrating fish was considerably smaller 
than Halttunen et al.’s (13 vs. 166 individuals), the fish 
spawned artificially in a hatchery rather than naturally 
in the river, and our outmigrating fish were all females 
(as opposed to mixed sex in Haltunnen et al. (2013)). 
