284 
Fishery Bulletin 11 6(3-4) 
Table 1 
Minimum, median, and maximum values of total length (cm), mass (g), and body condition 
index for kelt Atlantic salmon (Salmo salar) released at an upriver and a downriver site on the 
Penobscot River, Maine, on 17 November 2015. 
Downriver 
Upriver 
Min 
Med 
Max 
Min 
Med 
Max 
Total length 
72 
76 
85 
73 
77 
81 
Mass 
3900 
4481 
5901 
3338 
4454 
5561 
Body condition index 
-0.03 
0 
0.13 
-0.13 
0 
0.16 
Fish capture and tagging 
Throughout the 2015 upstream migration, adult Atlan¬ 
tic salmon were collected at the Milford Dam lift and 
trap facility (rkm 61) by personnel of the Maine Depart¬ 
ment of Marine Resources. After collection, the fish were 
transported to Craig Brook National Fish Hatchery, ap¬ 
proximately 56 km south of the dam, on a different river 
system. There, fish were held until they spawned in the 
fall. During the spawning process, fork length, mass, 
and sex of each fish were recorded. We used the residuals 
of log(Mass) regressed on lo giTotal length ) to calculate 
body condition (Halttunen et al., 2013). On November 
17, 2015, after several days of recovery from the arti¬ 
ficial spawning procedure, 55 fish were anaesthetized 
with tricaine methanosufonate (MS-222) and surgically 
implanted with MCFT2-3L 5 radio transmitters (Lotek 
Wireless, Inc., Newmarket, Canada). A 16-mm incision 
was made offset from the ventral line and posterior 
to the pelvic fin girdle on the ventral side of the fish 
to enable closure with two or three sutures of Vicryl 
size 4-0 (Ethicon, Inc., Somerville, NJ). Tags had an 
air mass of 25.0 g and a water mass of 11.0 g, ensuring 
that they were <1% of the estimated body mass of the 
study fish, which ranged from 3.3 kg to 5.9 kg (median 
mass=4.4 kg). Transmitters had a 2.5- second burst rate 
and a tag life of < 170 days because of their use in a 
previous study. Of the 55 fish, 27 were released at the 
boat launch in Brewer, Maine (rkm 43.5, a 26-mile drive 
from the hatchery ), downstream of the head of tide, and 
28 were released at the boat launch in Passadumkeag, 
Maine (rkm 92, an 85 km drive from the hatchery). 
All releases took place within five hours of surgery on 
November 17, 2015. 
Radio tracking 
We monitored fish movements over a seven-month peri¬ 
od from mid-November 2015 to mid-June 2016, primar¬ 
ily by mobile tracking with a portable receiver (SRX 
800, Lotek Wireless, Inc.) during trips in the study 
5 Mention of trade names or commercial companies is for iden¬ 
tification purposes only and does not imply endorsement by 
the National Marine Fisheries Service, NOAA. 
area. These trips took place between Orrington (rkm 
33) and Medway, Maine (rkm 149), on a weekly basis; 
more frequent trips (two or three/week) were made be¬ 
fore ice-in (late December) and after ice-out (late April). 
Once a month, we used mobile tracking along the Pis¬ 
cataquis River to detect any fish that moved upstream 
through the Howland Dam bypass. All mobile tracking 
was conducted by using cars with omnidirectional an¬ 
tennas traveling along the side of the river on ME-116, 
U.S. Highway 2, and ME-178 at an average distance 
of 0.332 km (min 0.037 km, max 0.965 km), and an 
average speed of 64 km/h (40 mph). We also employed 
an array of three stationary radio receivers (SRX-DL, 
Lotek Wireless, Inc.) equipped with unidirectional Yagi 
antennas that were located between Orrington (rkm 
33) and Indian Island, Maine (rkm 62; Fig. 1). One 
aerial survey, conducted aboard an airplane with a por¬ 
table receiver, covered the mainstem Penobscot River, 
the Penobscot estuary, and the Piscataquis River on 
January 25, 2016. 
Because a study with similar tagging methods 
found no tag loss in Atlantic salmon smolts (Stich et 
al., 2014), we assumed no tag loss. However, fish that 
remained in the same position throughout the winter, 
spring, and into the summer of 2016 were presumed 
dead. We used x 2 tests to test for differences in sur¬ 
vival, successful migration, and behavioral differences 
between the upper and lower release groups. Only fish 
that were detected at the most downstream receiver 
or fish that were detected in-river over the duration of 
the study were included in the survival analysis. Wil- 
coxon-Mann-Whitney tests were used to compare body 
condition index between groups of fish with different 
movements or survival outcomes. For the purposes of 
this study, we considered overwinter survival to take 
place between November and late April (ice-out); any 
movements after this were considered to take place 
during the spring, and we stopped tracking fish in 
early July 2016. 
Results 
We collected and tagged 55 fish from the hatchery: 53 
females and two males (Table 1). All were multiseawin- 
