MOVEMENT, HABITAT USE, AND DAILY ACTIVITY PATTERNS
OF TROPHY BROWN TROUT
IN THE SOUTH BRANCH OF
THE AU SABLE RIVER, MICHIGAN
by
David FL/Clapp
/
.l/'
A Thesis Submitted in Partial Fulfillment of the
Requirements for the degree of
Master of Science
School of Natural Resources
The University of Michigan
1988
Committee members
Associate Professor James S. Diana, Chairman
Adjunct Professor William C. Latta
BX§OffiCiQ examiner Richard D. Clark, Jr., Ph.D.
Acknowledgments
I would like first to express my sincere appreciation to
the members of my committee; without their work and advice
this project would not have been possible. Dr. Latta
arranged for office space, equipment, and computer use
through the Michigan Department of Natural Resources, as well
as providing constructive comments and criticisms throughout
the course of the study. Dr. Diana arranged all of the
funding for this project, and provided criticism and
direction whenever needed. Additionally, Dr. Diana's
practice of treating his students as friends and peers made
my studies at the University of Michigan truly enjoyable and
rewarding. Dr. Clark was the driving force behind all
aspects of this project; without his continued help and
encouragement during my undergraduate and graduate work, this
study would not have been completed. His discussions of
numerous ideas added considerably to the following document.
Major funding for this study was provided by the
Michigan Fly' Fishing Club, based in Livonia. The club
members showed tremendous interest in the project; they
provide a good example of the positive role that trout
fishing organizations can and must play in the conservation
of this valuable resource. I would especially like to thank
Bill and Fran Merrill. Bill provided a great deal of help to
me in the field, and made constructive comments on early
drafts of this thesis. Both Fran and Bill showed me
ii
tremendous hospitality during my two summers in Grayling.
I received technical advice and support from a number of
people. From the University of Michigan, I would like to
thank Professors Gary Fowler and Charles Olsen. Everyone at
the Michigan Department of Natural Resources Institute for
Fisheries Research in Ann Arbor was extremely helpful. I
would especially like to thank James Gapczynski, who helped
with much of the field work, made slides for numerous
presentations, and constructed necessary equipment for work
in the field. Roger Lockwood, Al Sutton, Barbara Lowell,
Barbara Gould, and Grace Zurek all helped immeasureably.
I would like to thank the Michigan State University
Fisheries Department and Land Management Office for providing
lodging at their Wa Wa Sum Lodge Research Facility in
Grayling, Michigan. Kevin Gardiner, the resident manager at
this facility, helped on numerous occasions with field work
and was a valuable friend during my two summers in Grayling.
A number of friends provided support and encouragement
to me during my years in Ann Arbor. I would like to
especially thank Steve Dater, Dave Lucchesi, Shelly Gray,
Lynn Jolicouer, and Jill Ciolli.
Finally, I would like most of all to thank my parents,
Richard and Martha, and my sister, Anne, for all they have
done. I owe them a debt of love and support that is
impossible to repay.
iii

Table _§ Contents
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. ii
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. v
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. vii
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. viii
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 1
Methods
Study Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 6
Implant of Transmitters . . . . . . . . . . . . . . . . . . . . . . .. 8
Location of Fish . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 11
Movement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 13
Habitat Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 14
Daily Activity Patterns . . . . . . . . . . . . . . . . . . . . . . .. 16
Results
Implant of Transmitters . . . . . . . . . . . . . . . . . . . . . . .. 19
Movement . . . . . . .._ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 20
Habitat Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 24
Daily Activity Patterns . . . . . . . . . . . . . . . . . . . . . . .. 32
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..v . . . . .. 43
Management Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 52
Literature Cited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 54
iv
List Qf Figures

Figure
Site use by fish #1,
South Branch of the Au Sable River, showing
various landmarks in the study area . . . . . . . . . ..
Range of movement by eight brown trout—
tracked during the present study . . . . . . . . . . . ..
Frequency of use for 57 quadrats occupied by
brown trout in the South Branch of the
Au Sable River . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Summer site use by fish #6, tracked from
June to December, 1987 . . . . . . . . . . . . . . . . . . . . . . ..
tracked during summer
1986 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Site use by fish #2, tracked during summer
1986 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Summer site use by fish #7, tracked from
June 1987 to May 1988 . . . . . . . . . . . . . . . . . . . . . . . ..
Comparison of available mean water velocity to
that used by three brown trout tracked during
summer 1987 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Comparison of available bottom water velocity
to that used by three brown trout tracked
during summer 1987 . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Page
22
23
25
26
27
28
29
30
10.
11.
12.
13.
14.
15.
16.
Comparison of available cover to that used
by three brown trout tracked during summer
1987 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Comparison of available substrate to that used
by three brown trout tracked during summer
1987 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Comparison of available water depth to that
used by three brown trout tracked during summer
1987 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Local activity pattern for brown trout tracked
during June . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Local activity pattern for brown trout tracked
during July . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Local activity pattern for brown trout tracked
during August . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Distributions of upstream (+) and downstream
(-) long—range activity . . . . . . . . . . . . . . . . . . . . ..
31
33
34
37
38
39
42
vi
List _§ Tables

Table Page
1. Summary of brown trout transmitter implants
performed between May 1986 and June 1987 . . . . .. 9
2. Range of movement and home site use by fish
tracked between May 1986 and May 1988 . . . . . . . .. 21
3. Mean monthly local activity for brown trout
tracked in summer . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 35
4. Mean and maximum long-range activity for fish
tracked between May 1986 and May 1988 . . . . . . . .. 41
vii
Abstract
Most previous studies of brown trout §almg_ trutta
ecology and behavior have focused on smaller fish, or fish in
lakes or under controlled conditions. Very little work has
been done in investigating the ecology of large, free—
ranging, stream—resident brown trout. The present study was
undertaken to monitor the movement, habitat use, and daily
activity patterns of these fish.
Radio transmitters were implanted in eight brown trout
between 437 mm and 635 mm from the South Branch of the Au
Sable River, Michigan over a two year period. Daily tracking
during summer (May — August), and tracking at two—week
intervals during fall and winter (September — April) was used
to determine movement, habitat use, and activity patterns of
these fish. Range of movement and home site use was
defined, and measured for each fish tracked. I evaluated
habitat use by comparing habitat in stream quadrats used by
fish to that available in quadrats chosen at random from
throughout the river. Two measures of brown trout activity
were defined; local activity and long—range activity. Local
activity was measured by counting fluctuations in radio
signal strength over a 24-h period, and long—range activity
was measured as the linear distance covered by a fish between
consecutive daytime resting locations.
Range of movement for eight fish tracked varied from
370 m to 33.4 km. The average range of movement in summer
was approximately 5,000 m, while the average range of
viii
movement in winter was approximately 12 km. Movement
appeared to be nonrandom: that is, fish used a few locations
often and were seen to return to these sites after movement
to other areas of the river. Fish tracked during summer
periods used as many as four home sites; the average
separation between these sites was 386 m.
Brown trout chose deep, slow areas with heavy log cover.
Significant positive electivity was seen for mean and bottom
water velocities less than 10 cm/sec, depths between 46 and
60 cm, areas of cover including overhanging branches,
vegetation, and logs, and areas with silt substrate.
Distinct peaks in local activity were observed during
two summer months. A major activity peak in June occurred at
2200 h, but in July this major peak shifted to 0500 h. No
distinct activity peaks were apparent in August. Light
intensity accounted for almost 29% of the variance in local
activity levels. Seasonal differences in local activity may
also have been related to changes in food availability or
temperature.
Long—range activity observed in summer was significantly
less than that seen during winter. Average summer long—range
activity was less than 300 m, while average winter activity
was greater than 3000 m. However, extensive nighttime
"foraging" activity in summer was much greater than any
reported in previous studies. No significant upstream or
downstream trends in long-range activity were observed once
fish took up residence in an area; however, many of the fish
tracked made a long movement to upstream areas in fall, then
ix
remained in these upstream areas over winter. Significant
positive correlations were seen for long—range activity with
volume discharge and average daily air temperature.
Significant negative correlation was seen between long—range
activity and groundwater levels.
During the present study, six of eight fish tracked
moved out of a catch-and—release section of the South Branch,
making them vulnerable to harvest in sections of the river
not covered by special regulations. However, four of five
fish tracked during the period of peak fishing pressure (May—
August) remained in this catch—and—release section.
Possibly, increased harvest of trophy fish in areas adjacent
to regulated areas could be counted as an additional benefit
of these quality fishing regulations.
Introdggtion
Studies of stream—resident brown trout Salmo trutta
have traditionally focused on movement, habitat use, and
activity patterns. The results of this research have in
many instances been disparate. Some researchers have shown
that brown trout move very little. Cobb (1933) found that
fish tagged in a number of Connecticut streams showed greater
displacement from the site of initial release than brook
trout Salvelinus fontinalis, but that the majority of
recaptures still tended to occur close to the areas in which
brown trout had been tagged initially. Allen (1951) reported
similar findings from a mark—and—recapture study of a New
Zealand brown trout population. Bachman (1984), using visual
observations, found that brown trout in Pennsylvania streams
remained close to a single location throughout their entire
lives.
Other studies, however, suggest that trout may move
extensively, especially during fall and winter. Schuck
(1943), using data from fish caught in weirs in combination
with tagging, found that brown trout showed little movement
during spring and summer, but could be seen to move several
miles upstream to spawn during October and November, and then
return to their original locations sometime during the
winter. Jenkins (1969) observed both a resident and
transient segment of brown trout populations in the mountain
streams of California. Resident fish generally displayed
long-term (50 days) position stability while the transient
segment of the population was characterized by frequent
aggressive displays and the absence of a settled social
structure.
Similar disparities can be seen in the results of
studies examining brown trout habitat use and activity
patterns. The variables thought to constitute ideal trout
habitat have been defined (White 1975, Raleigh and Duff
1980), but use of this habitat by trout in streams may vary
depending on factors such as lifestage (Raleigh and Duff
1980), activity (Gosse and Helm 1981), competition (Fausch
and White 1981), or threat (Bachman 1984). Many studies on
the behavior and feeding patterns of brown trout have shown
that activity may occur in distinct bouts, but the timing of
these bouts may vary considerably. Oswald (1978) recorded
three daily peaks in trout feeding activity using an analysis
of electromyogram rhythms. Elliott (1970) found midday and
evening peaks in feeding activity through an analysis of
brown trout stomach contents. Chaston (1969) found that
brown trout were most active between dusk and dawn in
laboratory experiments.
Disparities between the findings of these studies result
in part from differences in the methodologies used and in
the size of fish being studied. Mark-and-recapture studies
used to investigate movement can only provide limited
information. Recapture locations are sometimes known
accurately only to within one or two miles (Cobb 1933), so
distances moved by fish may be greater than those actually
reported. The activities of fish between marking and
recapture are not known, and dependence on angler efforts for
tag returns may bias results to include a relatively large
proportion of returns from popular fishing sections and
during peak fishing periods. Important fall and winter
movements may be missed. Weir captures can provide
additional data, but an investigator using weir data still
has no information -(other than direction) on where fish
originated, or what conditions at these original sites may
have triggered the movements observed. Direct visual
observations, while valuable, place strict spatial and
temporal limits on the study of a fish population. For
example, fixed shore stations can only be employed in cases
where fish are readily visible from above the surface of the
water. For this reason, fish which are often under heavy
cover may not be observed using this method (Rankin 1986).
Underwater (SCUBA) observations have added tremendously to
our knowledge of trout behavior and habitat use (Fausch and
White 1981, Cunjak and Power 1986), but may still introduce
bias by displacing fish from areas actually chosen.
Obviously, visual observations cannot be used to investigate
nighttime behavior without the aid of specialized optical
equipment.
The size of individual brown trout in the studies cited
above varied, but few studies specifically included larger
stream—resident fish as a part of the population under
investigation ( the average size of fish in these studies was
approximately 250 mm ). However, most recounted isolated
incidents in which larger brown trout exhibited behavior that
was strikingly different from that observed in the rest of
the fish population under study. Cobb (1933) reported on a
brown trout that may have moved 80 km downstream. Shetter
(1967) described movements by some large (>330mm) brown
trout of up to 65 km from the site of tagging. Jenkins
(1969) gave examples of large brown trout periodically
leaving areas of cover to roam through a stream section under
investigation; at times making attempts to forage on small
trout. This roaming behavior, while limited to movements of
less than 100 m, was not observed among brown trout of
smaller size classes. Bachman (1984) found that the size of
a brown trout's "home range" may decrease between age I and
age V; after this point, fish may adopt a roaming or
migratory lifestyle (Bachman 1982).
Large, stream—resident brown trout appear to show
patterns of behavior distinct from smaller fish, and some of
the common methods "of study may not be appropriate for
elucidating these behaviors. The use of radio telemetry
overcomes many problems related to the study of brown trout
ecology. Telemetry observations allow precise location of
fish for documentation of habitat use (Diana et al. 1977),
and allow movement and activity to be monitored continually
(Diana 1980, Mesing and Wicker 1986). Information on
location and behavior of trout at night and in heavy cover
can also be obtained. Telemetry is ideal for the study of
large fish (which are rarely studied using other
methodologies), as large fish are capable of carrying long—
lived transmitters without impaired swimming ability (Winter
1983).
The present study was undertaken to increase the current
base of knowledge concerning the movement, habitat use,
and daily activity of large, free—ranging, stream resident
brown trout. The specific objectives of this study were: 1)
to monitor seasonal movement and habitat use by trophy brown
trout in the South Branch of the Au Sable River near
Roscommon, Michigan; and 2) to evaluate daily activity
patterns of these fish during the main fishing season (May—
August). In this study, a trophy fish was defined as any
brown trout over 432 mm in length.
Methods

This study was conducted on the South Branch of the Au
Sable River, near Roscommon, Michigan. The South Branch is a
coldwater river extending from Lake St. Helen in Roscommon
County to its confluence with the Mainstream of the Au Sable
River in Crawford County, approximately 25 km east of
Grayling, Michigan. Six 'dams along the length of the Au
Sable Mainstream prevent migration of brown trout from Lake
Huron.
The South Branch recieves a stable groundwater input,
and is supplemented by inputs from a number of feeder creeks;
most notably Robinson Creek, just upstream from the town of
Roscommon. The upper reaches of the river flow through low
lying swampy areas and the primary fish present are northern

pike Esox lucius, YEIlOw perch Perca flavesceng, suckers
Catostomus spp, and minnows (Cyprinidae) (Shetter 1967).
Quick discharge of rainwater through these low areas into the
river causes increased flow fluctuations in this upper
section of the river (Bosserman and Higgins _1969). Below
Roscommon, the river cools, the gradient becomes somewhat
steeper, and trout species, usually brown trout and brook
trout, dominate (Shetter 1967). I worked primarily on the
area of the South Branch between Roscommon and Smith Bridge
(Figure 1).
Mean annual flow on the South Branch, measured at Smith
. 83' I | AU
19,.
SnflH1Bndge
. Downeys \\\\\‘
Castle \ N
Marlbar
Chase Bridge \
Deerheart Valley Rd. \
Steckert Bridge ___>
Mead's Landing -—--"'" Kilmer Rd. 0 5
3&%; ‘/// I! II II


ROSCOMMON c KILOMETERS
8.x
_C CD
.5@
£0
Lake
St. Helen Figure 1. South Branch of the Au Sable River, showing
various landmarks in the study area. Fish for implant Of
transmitters were taken between Chase Bridge and the Castle.
Statewide trout regulations were in effect upstream of Chase
Bridge downstream to one km below the
Bridge. From Chase
Castle, a flies-only, zero creel limit regulation was in
effect. Downstream from this point to the Mainstream, a
flies—only, 5 fish creel limit regulation was in effect.
3
Bridge, is approximately 6.5 m /sec (Coopes 1974). Average
width of the river is approximately 20 m, and average
gradient is 0.09% (Shetter 1967). Average minimum water
temperature in July is approximately 21 C upstream of
Roscommon, and approximately 18 C at Smith Bridge. Average
maximum water temperature in these same areas is
approximately 26 C and 24 C, respectively (Coopes 1974).
This temperature distribution is somewhat unusual. Many
streams, particularly those in mountainous areas, have cool
upstream reaches and warmer downstream reaches.
Implant Qf Transmitters
Transmitters were successfully implanted in eight brown
trout between May 1986 and June 1987. Another sixteen
implants were unsuccessful (Table 1). Fish for transmitter
implants were taken using D.C. electrofishing gear. An
incision was made into the abdominal cavity either through
the lateral body wall or ventrally, anterior to the pelvic
girdle. A transmitter was inserted through this incision,
and the incision was closed using non-dissolving nylon
sutures. Fish were released into the river near the site of
capture.
Transmitters used (from Custom Telemetry and Consulting;
Athens, Georgia) were approximately 4 cm long, 2 cm in
diameter, and weighed approximately 12 g (1986) or 20 g
(1987). Transmitter batteries used in 1986 failed earlier
than expected, and were replaced with larger batteries in
Table 1. Summary of brown trout transmitter implants
performed between May 1986 and June 1987. For Days tracked,
NS indicates fish tracked successfully for less than two


weeks, and 0 indicates fish dying immediately following
surgery.
Date Temper- Length Weight Days Implant
implanted ature (C) (mm) (g) tracked site
May 5, 1986 -- 488 1112 102 Lateral
432 810 NS Lateral
396 640 NS Lateral
396 600 NS Lateral
391 570 0 Lateral
396 610 NS Lateral
437 830 52 Lateral
Jul 3, 1986 21 470 1200 0 Lateral
546 1670 0 Lateral
541 1710 0 Lateral
500 1190 0 Lateral
Oct 23, 1986 10 589 1850 80 Lateral
521 1550 114 Lateral
569 1570 0 Lateral
635 2050 93 Lateral
627 2500 NS Lateral
Table 1 (continued)...


Date Temper— Length Weight Days Implant
implanted ature (C) (mm) (g) tracked site
May 5, 1987 11 470 ———— NS Lateral
485 ———— NS Lateral
462 ———— NS Lateral
419 —-—- NS Lateral
Jun 3, 1987 18 505 ———— 346 Ventral
488 ———— 345 Ventral
510 ———— 199 Ventral
Aug 8, 1987 20 556 -—-— 0 Ventral

10
1987. Each transmitter was encased in surgical wax and bore
identifying information, including a telephone number to be
contacted in the event one of the fish was found dead, or
captured by an angler. Each fish was tracked on a different
frequency between 49 and 50 MHz.
Location g§_Fish
During the summer months (May — August) I attempted to
locate every fish on each day. When fish moved to widely
separate areas this was not always possible. However, I
usually did not go more than two days between locations of a
given fish. During the winter months (September — April),
fish were located at approximately two—week intervals. Data
collected during the two weeks immediately following implant
of transmitters were discarded, since fish often exhibit
erratic behavior during this time (Mesing and Wicker 1986).
I initially located fish by floating the river in a
canoe, using a scanning receiver (Model - Challenger 200) and
60 cm loop antenna (both from Advanced Telemetry Systems;
Isanti, Minnesota). Using this system, I could detect fish
at a distance of approximately 200 meters. If a fish had
been located recently, I began searching at the site where I
had last found that fish. After a radio signal was detected,
I took bearings from a number of positions on the river bank,
upstream and downstream of the estimated position of the
fish, in order to more accurately determine that position. A
smaller (15 cm) loop antenna was often used for close—range
11
work. When I had determined a fish's location, I waded into
the river to verify the position. I could generally approach
to within five meters (sometimes to within one meter) of a
fish before it was disturbed, so that verification was by
sight of the fish, by maximum signal strength, or in some
cases by maximum signal strength followed by a drop in
strength (indicating rapid movement by the fish). Fish were
only disturbed to make habitat measurements, and generally
they remained near or in cover at the position of location,
even as habitat measurements were taken. I assumed that
these disturbances did not significantly influence the
movement and activity patterns observed.
The locations of all fish were plotted on maps
constructed from aerial photographs and topographic maps of
the area. On these maps, the river was divided into quadrats
10 meters long ( thalweg distance ) with a width equal to
either 1/2 the width of the river, or equal to 10 m (if the
river was wider than 20 m ). The size of this quadrat
represents an area large enough to include that entire
portion of the river in use by a trout at the time of
location, and small enough to allow for accurate location and
habitat data collection.
In analyzing movement and activity, the data I collected
were separated into two groupS. The first group included
summer data, collected between May 1 and August 15; the
second group included fall/winter data, collected between
August 16 and April 30. This second group will hereafter be
referred to as winter data. Two fish were tracked
12
exclusively during a summer period (May - August 1986), three
fish were tracked exclusively during a winter period (October
1986 — February 1987), and three fish were tracked through
both summer and winter periods (June 1987 — May 1988). This
division into summer and winter was not an arbitrary one, but
rather based on observations made in the field. Fish tracked
during this study tended to exhibit much more extensive
movements beginning in mid—August.
Movement
Range Of movement is defined in this study as the
distance between extreme upstream and downstream telemetric
locations of each fish. Home sites within this range were
identified for fish tracked during summer periods, and I
compared the average separation between home sites for each
of these fish. A home site was defined as: l) a quadrat
that was used five or more times by a given fish, or 2) a
quadrat which a fish returned to after moving to a separate
quadrat in the river (thus exhibiting a homing tendency). Two
adjacent quadrats which both fulfilled the requirements of a
home site were considered to be a single home site. I also
plotted distributions of quadrat use for each fish tracked
during a summer period, and compared the average number of
days that each fish remained at a quadrat before moving.
13
Habitat Use
Quantitative habitat data was collected exclusively for
fish located during summer tracking. This included two fish
in 1986 and three fish in 1987. During summer 1986, I
measured depth, mean water velocity, substrate type, and
cover type at a single point that I took to be the focal
location of the fish being tracked. There were two major
problems associated with this methodology; 1) radio
locations were not accurate enough to be taken as precise
focal point values, and 2) because large brown trout were
found to move a great deal (as compared to smaller fish),
measurements at a single point were not totally
representative of habitat being used. To correct these
problems, I made some modifications for work during summer
1987. Results reported in this study reflect only habitat
data collected in sites used by fish tracked during 1987.
After locating a fish and identifying which map quadrat
encompassed its position, I completely characterized the
habitat present in that quadrat. Eighteen quadrats used by
fish were characterized. Fish often used a quadrat more than
once; multiple measurements were taken on six of these
eighteen quadrats. I established transects at the upstream
edge, middle, and downstream edge of each quadrat, and
determined water depth, mean and bottom water velocity,
substrate, and cover type present at one meter intervals
along each of these transects. These were assumed, from a
review of the available literature, to be the five most
14
important habitat variables. Approximately thirty
measurements were made in each quadrat, depending on stream
bank morphometry (3 transects x 7—10 m per transect). Mean
water velocity (at 0.6 of the depth of the water column) and
bottom water velocity (from 1—5 cm above the substrate) were
measured using a Swoffer 2000—1 Open Stream current meter.
Predominant substrate type at each meter interval was
estimated by sight as belonging to one of five categories —
silt, sand, gravel (< 2 cm), small cobble (2—10 cm), or large
cobble (> 10 cm). Predominant cover type at each meter
interval was estimated by sight as belonging to one of six
categories - logs, brush, vegetation, boulders, overhang, or
open. Logs provided instream cover for fish, and included
trees, limbs, boards, and combinations of these items with
individual widths or diameters greater than 10 cm. Brush
also provided instream cover for fish. but included no
individual items with widths or diameters greater than 10 cm.
This category included primarily tree tops which lay into the
water, as well as some flooded riparian vegetation.
Overhanging cover provided no instream shelter for fish, but
acted only to shade fish from direct sunlight and overhead
disturbance.
In addition to quadrats enCOmpassing brown trout
locations, habitat was characterized for nineteen other
quadrats chosen at random from within the stream sections
used by radio—tagged fish. I used Strauss' (1979) linear
index of electivity to compare habitat use data pooled from
three fish tracked in 1987 to data from these random
15
quadrats. This index ranges from -1 to 1, with positive
values indicating preference and negative values indicating
avoidance or inaccessibility. I used a 5% level of
significance (t—test, P=0.05) to test whether selection by
brown trout for five habitat variables was significantly
different from zero.
Only habitat data collected in the five meters of each
quadrat closest to the stream bank was used for the analyses
reported here, since this was where fish were located in all
cases. Analyses using data from the entire quadrat gave
similar results.
Daily Activity Patterns
Two measures of activity were obtained during the
present study. The first, local activity, reflects fish
activity in a limited area within range (approximately 200 m)
of a radio reciever and antenna mounted on the stream bank.
Observations to determine local activity patterns were made
throughout the summer (June — August) during 8 three-hour
(1986) or 12 two-hour (1987) observation periods, randomized
over a day. A complete 24-h cycle was covered approximately
once every two weeks. Two fish were used for observations in
1986, and an additional two fish for observations in 1987.
If possible, fish were observed in sequence, one or both fish
each day. If the "target" fish could not be located before
the observation period in question, the second fish, if it
could be located during that period, was used for the
16
observation.
During these observations, I used a tripod to mount an
antenna on the bank of the river within signal range of the
fish. It was not necessary to know the exact location of the
fish, because I was only monitoring absolute activity levels.
I was able to detect changes in the distance from and
orientation of a fish to the mounted antenna as fluctuations
in radio signal strength. The number and magnitude of these
fluctuations reflected three levels (resting, turning, or
continuous) of local activity. I determined that
fluctuations were caused by fish activity and not simply
background "noise" by observing the signal response on a
chart recorder to a transmitter placed in a dead fish which
was moved manually to simulate swimming activity.
During each observation period, I recorded the number of
fluctuations in radio signal strength over a one—minute
interval, once every five minutes. Local activity in this
study is defined as the number of fluctuations in signal
strength per minute of observation. Measures of local
activity for fish were stratified by month and year, because
I assumed that this activity would vary in response to
seasonal changes in day length, water temperature, and food
availability. Differences in local activity levels between
these strata were investigated using a two - way analysis of
variance (ANOVA). I plotted mean hourly local activity
levels for each month and year to show 24-h patterns, and
attempted to relate the observed variations in activity to
metabolic scope for activity, light intensity, and feeding
17
rate using a regression analysis. Metabolic scope for
activity (the difference between maximum (Q ) and standard
max
(Q ) metabolism at a given temperature) and feeding rate
stand
(in g/hour) were calculated using water temperature data
recorded during the present study, and the following
equations from Elliott (1975, 1976):
Scope = Q - Q , where
max stand
(B ) ((B ) x T)
Q = A X W 1 x e 2 , and
max 1
(B ) ((B ) x T)
Q = A X W 3 x e 4
stand 2
(D x T)
Rate = C x e
A , A , B , B , C , C , D , and D terms are temperature
1 2 1 2 1 2 1 2
dependent constants determined by Elliott (1975, 1976), T is
ambient (water) temperature, and W is the fish's weight in
grams. Light intensity was included in the regression model
as a linear function with a minimum at midnight (0000 h) and
a maximum at noon (1200 h).
The second activity measure, 19n9-range activity, was
defined as the linear distance, upstream (+) or downstream
(-), covered by a fish between consecutive daily resting
lOcations. Yearly and seasonal variations in long—range
activity, as well as differences in long—range activity
18
between individual fish, were evaluated using a nested ANOVA,
or, in some cases, a Kruskal—Wallis test. Frequency
distributions of upstream and downstream long—range activity
for summer and winter periods were compared using a
Kolmogorov—Smirnov test. I also calculated correlations
between long-range activity and volume discharge, daily
Change in VOlume diSCharge, groundwater level, daily average
air temperature, daily high and low air temperatures, and
day length. Discharge and groundwater data were obtained
from the United States Geological Survey station in Grayling,
MiChiQaR (CraWford CQuntY), and the remaining data were
obtained from the National Weather Service station at
Roscommon County Airport, Houghton Lake, Michigan. I used a
5% level of significance (P = 0.05) for all regression
analyses and statistical tests.
. Results.
ZEELQQL Qi Transmitters
Mortality of brown trout following surgical implant of
transmitters was relatively high (67%). Mortality seemed
to depend on implant site, as well as water temperature at
and following the time of imPlant. Fish for which
transmitters were implanted through incisions in the lateral
bOdy wall experienced extremely high mortality (75%).
Lateral incisions may have severed connective tissue, causing
delays in healing and increasing the risk of infection. Fish
19
with transmitters implanted ventrally showed much lower
mortality (25%). The only mortality following a ventral
implant occurred in August 1987, when water temperatures
reached 20 C (Table 1). It was apparent that the majority
of mortalities occurred during periods of warm water,
regardless of implant site, and I attempted to avoid these
periods in later surgery.
Movement
Range of movement varied considerably among individual
fish. The smallest range was 370 m, and the largest was 33.4
km (Table 2). No seasonal trends in range of movement were
apparent. The average range of movement in summer was 4.9
km, while the average range of movement in winter was 11.9
km. Six of the eight brown trout tracked moved out of the
catch—and—release section and into areas of the river under
statewide trout regulations (Figure 2). However, four of
five fish tracked during summer periods remained in or near
this section of river throughout most of the summer.
Movement by fish, especially in summer, appeared to be
nonrandom. Trout used a few locations often, and returned to
these sites after movement to other areas of the river. Of
fifty—seven different quadrats used by trout during this
study, eight were used seven or more times (Figure 3).
Quadrats used by fish tracked in 1986 were also used on three
occasions by fish tracked in 1987. Overlap in site use
20
Table 2. Range of movement and home site use by fish
tracked between May 1986 and May 1988. Mean separation
refers to average distance between home sites used by a given


fish. Overall values are averages of values for individual
fish. Standard deviation in parentheses.
Fish Range Number of Mean Percent time
number (m) home sites separation in home sites
§2EE2£
1 670 4 40 (22) 86
2 370 4 63 (37) 94
6 16,800 1 —- 97
7 1,900 4 1,053 (732) 88
Overall 4,935 (7,938) 386 (628) 91
Winter
3 470 l -- 33
4 11,610 1 " 25
5 2,110 0 “' ~—
8 33,420 1 '- 40
Overall 11,902 (15,162) -- 33

21
I
Kilmer Road 40 - .
"I
r-—--1
H
Roscommon
30 -
Chase Bridge -— ---------------- -._


0+1
_ 2
High Banks - ------------------------------------ _.._,__
River Landmarks and Disiance (km)
10 -
Smiih Bridge
I


Mainsiream 0 = = = s s : =
1 2 3 4 5 6 7 8
Fish Number
Figure 2. Range of movement by eight brown trout tracked
during the present study. The catch—and—release (flies—only,
zero creel limit) fishing area is shown with horizontal
dashed lines. Upstream and downstream limits of fish
movements are indicated by high and low horizontal bars.
Vertical solid lines indicate summer range of movement,
vertical dashed lines indicate winter range of movement.
22










35 -)-
30 -P
25 --
:5
3
g- 20 --
I:
U-
2".
LL-
15 ~-
10 .-
5 air
, inl |.. .. m. .|.....l,. .I . ..|..........|l||.....|
Mainstream Kilmor Rd
Figure 3. Frequency of use for 57 quadrats occupied by
brown trout in the South Branch of the Au Sable River. Joyce
Kilmer Reed marked the upstream extent of movement, the
Mainstream marked the limit of observed downstream movement.
Quadrats used were not evenly distributed (in distance)
between these two limits.
23
occurred between two fish tracked in the winter of 1986 and
between two fish tracked during 1987—1988.
The number of home sites used by fish tracked during
summer ranged from one to four. One fish used a single home
site (Figure 4), while three others each used four home sites
(Figures 5, 6, and 7). The average separation between home
sites for these three fish was 386 m, and these home sites
were in use between 86% and 97% of the times fish were
located (Table 2). Fish tracked in summer moved between
quadrats (not necessarily home sites) about once every three
days (mean = 3.15 days), but one fish was observed in a
single quadrat for fifty—two consecutive days without
displacement.
Habitat Use
Fish appeared to select areas with mean and bottom water
velocities less than 10 cm/sec. In quadrants used by fish,
70% of the measures of mean velocity were less than 10
cm/sec, while only 46% of mean velocity measures in random
quadrants were less than 10 cm/sec (Figure 8). The
corresponding values for bottom water velocity were 79% and
56%, respectively (Figure 9). Electivity indices were
significantly greater than zero for both variables (P <
0.05).
Fish consistently used areas with overhang, vegetation,
and log cover (Figure 10). Significant positive electivity
was seen for all three of these cover types (P < 0.05), while
24
Frequency of Siie Use (n)




30 ' 40
Dlsfance from Mainstream (km)
Figure 4. Summer site use by fish #6, tracked from June
to December, 1987. Home sites, as defined in text, are
indicated with an "*".
25
121-
s
10..
A
c:
v
3 3..
3
CD
7:
(f)
"'5 s--
>
O
C
Q)
:3
U-
m 4.-
he
LI..-
4:





Distance from Mainsiream (km)
Figure 5. Site use by fish #1, tracked during summer
1986.
26
I.
w --
I
5
l
E a: 2w 2 ass»;
I I I I I I I I
I I I I



5
23.
Disiance (ram Mainsiream (km)
summer
tracked during
use by fish #2,
6. Site
Figure
1986.
27
.0? \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ \\\ s
.60. \\\\\\\\.\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\>
\\\\\\\
.0.- S\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
Q\\\\\.
.0.- \\\\.\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\>
I
(a --
a
s
4
2
a

. E a: em 2 55:8:
26
25
24
Distance from Mainsiream (km)
tracked from June
Summer site use by fish #7,
Figure 7.
1987 to May 1988.
28
Used
[:1 Available
Frequency (Z)






g. ;§Izmi_i:m_l: l—l:

S asses
x\‘\§®¥‘¥®

Mean Waler Velociiy (cm / sec)
Figure 8. Comparison of available mean water velocity to
that used by three brown trout tracked during summer 1987.
Velocities for which use was significantly different
from
that available (P<0.05) are indicated with an "*".
29
80s






Used
[J Available

Bailom Waier Velacily (cm/sec)
10 --
a
60 #-
501.
E (0-
5‘
5
0?; 30 -
(I:
20 an
10 -
ll : -
h B
%' §;‘ Qgb
Figure 9.
Comparison of available bottom water
velocity
to that used by three brown trout tracked during summer 1987.
Velocities for
which use was significantly different from
that available (P<0.05) are indicated with an "*".
3O









#
50.. -—
50 .. h ‘ *
Used
40.. [:1 Available
5
>_
g 30-
Q)
:3
5
L5 1
20 *
10. gr) *
n : a: m:§ :§_i: :
as? my a} gs we
at we
Caverlype
Figure 10. Comparison of available cover to that used by
three brown trout tracked during summer 1987. Cover types
for which use was significantly different from that available
(P<0.05) are indicated with an "*".
31
significant negative electivity (avoidance) was seen for open
areas (P<0.05). Fish used predominantly silt and sand
substrate (Figure 11), but significant positive electivity
was seen only for silt ( P< 0.05). Selection was also seen
for depths between 31 and 60 cm, and depths greater than 75
cm (Figure 12), but electivity was only significantly greater
than zero for depths between 46 and 60 cm (P < 0.05).
An interaction between habitat variables was apparent.
Significant correlations were seen between water depth and
mean water velocity, water depth and bottom water velocity,
water depth and substrate, mean water velocity and substrate,
and bottom water velocity and substrate for data from
quadrats used by fish ( P (0.05; r = 0.28, 0.15, 0.36, 0.60,
and 0.43, respectively).
Daily Activity_Patterns
Average local activity (signal fluctuations/min of
observation) was not significantly different between 1986 and
1987. However, there were significant differences in
activity between months within a given year (Table 3). In
both 1986 and 1987, local activity in August was
significantly greater (P < 0.05) than activity in’ June or
July. There were no significant differences between activity
levels in June and July of either year.
There were significant differences between hourly,
local activity levels within a given month. In June, major
peaks in local brown trout activity were observed at 0100,
32
"w
.\\B Used
[:1 Available















\\\\\\\\\\\\\\\\\u





\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\u

F
I
w
P
I
0
3

I
q
0. o 0
g 5:23:
a
a" at a)“ as"
Subslrale lype
as,
11. Comparison of available substrate to that used
Figure
Substrate
racked during summer 1987.
brown trout t
by three
which use was
ent from that
significantly differ
for
types
ated with an "*“.
are indic
available (P<0.05)
33
Used









(:1 Available
5
>
O
c
Q)
:3
U“
22
Lb
R
g
. : m :m m
s s s a e
’\‘ a. 6" at "
Waler Deplh (cm)
Figure 12. Comparison of available water depth to that
used by three brown trout tracked during summer 1987. Depths
for which use was significantly different from that available
(P<0.05) are indicated with an "*".
34
Table 3. Mean monthly local activity for brown trout





tracked in summer. F/M = fluctuations in radio signal
strength per minute of observation, standard deviation in
parentheses. N is the number of observations.
June July August
F/M N F/M N F/M N
1986 5.01 585 4.73 375 6.30 238
(6.55) (4.45) (7.23)
1987 5.64 225 5.05 412 6.07 271
(6.58) (5.96) (6.14)
Overall 5.18 810 4.90 787 6.18 509
(6.56) (5.30) (6.67)

35
0500, 1430, and 2200 h (Figure 13). Evening (2200 h) was the
most active period of the day (approximately 13
fluctuations/min). The 0100 h and 0500 h peaks, at 8—9
fluctuations/min, were of approximately equal magnitude, and
secondary to the evening peak. The midday peak (1430) was the
third most active period of the day at 7 fluctuations/min.
In July, two peaks in local activity were observed, one
near midnight (0000 h) and another at 0500 h (Figure 14). The
morning peak (0500) was, at 21 fluctuations/min, the most
active period of the day in July, and was more than two times
greater in magnitude than the June morning activity peak.
The midnight activity peak, at 12 fluctuations/min, was
comparable in magnitude to the June evening peak.
In August, no distinct peaks in local activity were
apparent (Figure 15). A low point occurred at 1100 h, when
activity dropped to 2 fluctuations/min. Before and after
this low point, fish seemed to alternate between periods of
high and low activity every three to four hours. High points
ranged between 6 and 11 fluctuations/min, and low points
between 3 and 5 fluctuations/min.
A linear regression model including metabolic scope,
light intensity, and feeding rate accounted for 29% of the
observed variance in heurly local activity (P<0.05). Light
intensity accounted for the greatest proportion of this
variance.
Long-range activity (the distance covered between
consecutive daytime resting locations) of fish tracked during
summer was significantly less (P<0.05) than that of fish
36
lulllilllllI... I
llalltllllnlull
‘1
\





2538:9185 £52
2300
1200 (800
f Day
Time 0
Local activity pattern for brown trout tracked
Figure 13.
Maximum
30
activity
possible
June.
during
Solid
fluctuations/min.
represents mean activity
line
calculated by combining observations for four separate fish,
95%
Dashed lines indicate
two from 1986 and two from 1987.
confidence intervals.
37
30"
25"



Activity (fluctuations/min)





0
0000 0000 (200 1800 2300
Time of Day
Figure 14. Local activity pattern for brown trout tracked
during July. Solid line represents mean activity calculated
by combining observations for three separate fish, one from
1986 and two from 1987. Dashed lines indicate 95% confidence
intervals.
38

. - Q l I l l I. l l l l l l . . l l l l l l
‘ ‘ ‘ ‘ ‘ I ‘ ‘ ' ‘ i t ‘ ' I I r ' ' ‘, r ‘

8 086420
1|


25225285 3.52.
2300
1800
f Day
ime o
T
Local activity pattern for brown trout tracked
Figure 15.
during August. Solid line represents mean activity calculated
combining observations for three separate fish,
one from
by
Dashed lines indicate 95% confidence
1986 and two from 1987.
intervals.
39
tracked during winter (Table 4). Mean activity for all fish
tracked during summer combined was 239 m, while the mean
value for all fish tracked during winter combined was 3,103
m. However, sample sizes (number of observations) were low
for winter periods. Two fish tracked during summer showed
average long—range activity of less than 100 m, while one
fish showed average activity of almost 400 m. A fourth fish
showed average long—range activity of 480 m, but was located
in the same position for thirty—five consecutive sampling
periods. Three fish tracked during winter showed average
movements greater than 500 m, and this average movement went
as high as 11.2 km for one fish.
Seasonal distributions of upstream and downstream long—
range activity were found to be significantly different ( P <
0.05). The majority of summer movements were 50 m or less,
up or downstream (76%), while only 33% of winter movements
were within this range, and 64% were 100 m or more up— or
downstream (Figure 16). There were no significant upstream or
downstream trends in number of movements observed once fish
took up residence in an area; however, many of the fish
tracked made a long movement to upstream areas in fall, then
remained in these upstream areas over winter.
Significant correlations of long—range activity with
abiotic variables varied from year to year and between
individual fish. A regression of 1986 activity on average
daily air temperature, volume discharge, daily change in
volume discharge, and groundwater levels (using data for all
fish combined) was significant (P < 0.05), and explained
40
Table 4. Mean and maximum long—range activity for fish
tracked between May 1986 and May 1988. Standard deviation in


parentheses.
Fish Maximum Mean Number of
number - (m) (m) observations
Summer
1986 1 540 48 20
(120)
2 270 46 52
(69)
1987 6 16,800 480 35
(2,840)
7 1,895 399 34
(676)
Winter
1986—87 3 460 160 5
(219)
4 8,320 1,750 7
(3,102)
5 1,760 761 7
(670)
1987—88 8 22,730 11,220 5
(11,345)

41
Summer
[1 Winler












5
(I,
c:
:2
15
E:
(D
(I)
4:
c:
eels
Figure 16. Distributions of upstream (+) and downstream
(-) long—range activity. Data combined for eight fish
tracked between May 1986 and May 1988. Number of
observations was 141 in summer and 24 in winter. Numbers on
the abscissa indicate upper limits for inclusion in each
category.
42
2
approximately 25% of the observed variance in movement (r =
0.24). Average daily air temperature was the most
significant variable in this regression. Average monthly
long-range activity (all fish combined) was correlated with
average monthly air temperature, volume discharge, and
groundwater levels. A regression involving all three of
these variables was significant (P < 0.05) and accounted for
apgroximately 90% of the observed variance in movement
(r =0.91). Volume discharge accounted for the majority of
this variance.
Discussion
Trophy brown trout followed in this study had daytime
resting sites, but foraged away from these sites at night.
The frequency and extent of periodic movements and activity
away from and between resting sites was greater than any
previously recorded. The following general pattern became
apparent: trout showed limited local activity during the day
(corresponding to resting or turning activity), then
continuous local activity at dusk away from daytime resting
sites, sporadic or continuous foraging in midstream
throughout the night, and finally continuous activity back to
the previous resting site at dawn. This pattern was observed
for all fish tracked during the summer periods. One of these
fish, tracked in summer 1987, on numerous occasions moved
upstream about 1.5 km during the course of this nighttime
"foraging", then returned the same distance downstream to its
43
previous resting site the following morning. The behavior
described is in marked contrast to that exhibited by smaller
brown trout, which tend to maintain a single station of
limited area where they forage throughout the day (Jenkins
1969, Bachman 1984).
Summer movement and long-range activity by brown trout
was limited, as compared to winter movement and activity.
This observation is in general agreement with findings by
Cobb (1933) and Shetter (1967). The greater fall and winter
movement and long—range activity by fish observed in my study
was probably associated with spawning, but may have served
some other purpose as well. During November and December,
many of the brown trout in this study abandoned good spawning
habitat below Chase Bridge (where I observed approximately
80—90% of general spawning activity, including one instance
of a radio-tagged fish on a redd ) to move into what appeared
to be suboptimal spawning areas near and upstream of
Roscommon. Possibly these large brown trout were moving
upstream in search of better overwinter habitat. Cunjak and
Power (1986) indicated that space (habitat), and not food,
was probably a key resource for brown trout during winter
periods. Gosse and Helm (1981) attributed seasonal
differences in habitat use to changes in the amount of time
spent at various activities during winter periods. While
winter habitat use appeared, qualitatively, to be similar to
that in summer, the winter areas may have differed with
respect to water temperature, ice conditions, or food supply
— factors which were not monitored closely in the present
44
study. This question concerning the relative importance of
spawning versus overwinter habitat in determining fall
movements and winter site choice by brown trout is one that
needs further investigation.
There are a number of possible explanations for the more
extensive movements of large brown trout followed in this
study, as compared ’to those of smaller fish. Many fish
species are known to switch food types as they get larger.
This switch may necessitate movement over a larger area to
obtain this food (Dill 1978, Bachman 1982). Alexander
(1977) and Stauffer (1977) have shown that a switch from
insects to fish as the primary food items occurs for brown
trout in the Au Sable River at about 250—300 mm (age II or
III). Prey fish might be more abundant in the warmer,
upstream areas of the South Branch. Possibly the large brown
trout are moving to take advantage of these prey fish
populations when water temperatures allow them to penetrate
these areas. I
Some evidence exists to support the hypothesis that
temperature may influence movement of brown trout in the
South Branch. Monthly electrofishing surveys by Michigan
Department of Natural Resources (MDNR) personnel in the
catch-and—release section of the South Branch between Chase
Bridge and the Castle during summer 1987 found 6 brown trout
larger than 457 mm in this section in May, 14 in June, 20 in
(July, 18 in August, 12 in September, and 25 in October
(Richard D. Clark, Jr., MDNR, personal communication).
Temperatures in the upper reaches of the South Branch are
45
warmer than those in the catch-and—release section. Fish
numbers may increase in the middle of summer as a result of
fish moving from warmer upstream areas into this cooler area
of the river with greater groundwater input. Numbers may
increase in October as fish move into the catch—and—release
area to take advantage of good spawning habitat.
High positive correlation was found for long—range
activity with volume discharge, while high negative
correlation was found with groundwater levels. Caution
should be used in interpeting these results. Correlations
were obtained from monthly averages, and there were also high
correlations between the independent variables used. While
correlations and regression models cannot prove cause—and—
effect, the correlations obtained seem logical. Increases in
discharge might provide large brown trout with the perception
of security and cover, possibly leading to an increase in the
amount of fish activity and movement. Also, increases in
discharge can wash terrestrial food items into a river.
Movement may be greater due to increased feeding activity.
The Au Sable has one of the most stable flow regimes of
any trout stream in America (Richard D. Clark, Jr., MDNR,
personal communication); the relationship seen in this study
between brown trout movement and flow (volume discharge)
might be more important in streams where a typical seasonal
fluctuation in volume discharge is more apparent. Fish
movement might also be related to gradient and water velocity
in a given stretch of river. Faster water is harder to swim
against. Fish in a high gradient stream may use similar
46
physical locations as those used by fish in the Au Sable, but
the energy spent in maintaining and foraging from these
positions may preclude long range movements.
Fish tracked during the course of this study chose
deep, slow—velocity areas with heavy log cover. Habitat
utilized by brown trout was quite similar to that described
in the literature (Raleigh and Duff 1980, Gosse and Helm
1981, Shirvell and Dungey 1983). There were statistically
significant differences between areas used by these fish and
areas available to them, but the differences were not
extreme. This may be explained by the fact that the South
Branch is, on the whole, a stream with many areas of good
brown trout habitat along its length. The percentage of
usable habitat in the AuSable is probably 15% or greater, the
upper limit of the range given by Gosse and Helm (1981).
Habitat chosen by brown trout in a marginal stream would
probably show a greater contrast to habitat available to
them.
Habitat use is generally influenced by factors such as
activity (Stalnaker and Arnette 1976), competitive
interactions (Fausch and White 1981), and season (Cunjak and
Power 1986). Brown trout the size of those I studied are
probably the dominant stream—resident fish (Fausch and White
1981, Gosse and Helm 1981). For this reason, and due to the
low densities (3/ha: Richard D. Clark, Jr., MDNR, personal
communication) of these large fish in this section of the
river, I assumed that competitive interactions were not the
most important factor determining habitat use.
47
The effect of activity on habitat use was apparent in
this study. The quantitative data I collected on habitat use
was exclusively from daytime resting sites, but habitat use
during nighttime feeding and activity periods was obviously
different. These trout were often active at night in shallow
runs, or simply moved through a large variety of different
habitat types during periods of increased nighttime activity.
While quadrats used by fish and characterized in this
study probably provide a good description of the type of
habitat that is generally used by "trophy" fish, it is
important to remember that a diversity of habitat types is
necessary. Gosse and Helm (1981) indicate that age 0 brown
trout use macrophyte beds not used by juvenilles and adults.
Cunjak and Power (1986) saw significant age-specific
relationships for depth and velocity use: cover, on the other
hand, was used by all age classes in proportion to its
abundance in the environment. Appropriate habitat for all
ages of brown trout is critical to the maintenance of a
healthy fishery.
The number, timing, and size of the local activity
peaks exhibited by trophy brown trout changed seasonally from
June through August. The large evening peak in June activity
seen in this study may have been due to feeding activity by
brown trout on large mayflies ("Brown drake" Ephemera
simulans and "Michigan caddis" Hexagenia limbata) in mid— to
late June. However, subsurface feeding activity dominated.
I observed surface feeding by radio—tagged fish less than ten
times during the course of this study. Fish did not appear
48
to take up a feeding lane to feed on drift organisms, as
smaller trout do.
The peak activity period in July shifted to the morning
(0500 h), and was of greater magnitude than the June evening
peak. This shift may have occurred as a result of energetic
constraints related to the effect of water temperature on a
fish's metabolic efficiency. Daytime summer temperatures
above 20 C may have forced fish to feed primarily in early
morning, during the coolest portion of the day. While a
great deal of work has been done on behavioral
thermoregulation by fish species in lakes, the use of
behavioral thermoregulation by fish in river systems is an
area that warrants further investigation.
August local activity was more evenly distributed
throughout the 24—h period than that observed in other
months. This may have occurred as a result of higher water
levels in August 1986 and 1987, which could have afforded the
fish some security and prompted them to feed during parts of
the day when they generally would have been under cover. An
alternative explanation is that food may have been scarce in
August, and fish may have been forced to increase movement in
order to search for widely separate food items.
Swift (1962) found that the majority of brown trout
activity in lakes occurred during daylight hours, commencing
with a sharp rise in activity at dawn. Lab work by Chaston
(1969) indicated that fish were most active between dusk and
dawn during observations from spring through autumn. Oswald
(1978) found three daily peaks in feeding activity by
49
analyzing the electromyogram rhythms of brown trout feeding
in a lake, and indicated that these peaks of activity were
closely associated with photoperiod (dawn and dusk).
However, he also recorded some subsidiary activity "peaks
that appeared to be unconnected to photoperiod, and in
addition states that night feeding is probably a common
occurrence in brown trout. Elliott (1970) found midday and
evening peaks in feeding activity through an analysis of
brown trout stomach contents.
Just as movement may be site or river specific, so too
might activity. Patterns of activity may be based on food
distribution and temperature regimes, and these factors vary
depending on locality. Because of this, these findings of
site—specific and system—specific (lake versus river) brown
trout activity patterns should not be surprising.
Oswald (1978) indicated that trout activity patterns
may be influenced by variations in the rate of stomach
filling and evacuation, and Chaston (1969) stated that
pedator—avoidance may be an important consideration. In this
study, a model based on metabolic scope, feeding rate, and
light intensity explained some of the variation observed in
local activity, with light intensity being the most important
factor. While there may be some correlation between fish
activity and photoperiod, other explanations for fish
activity may be just as plausible. Temperature and food
availability have been shown to be correlated with light
patterns (Chaston 1969, Elliott 1970). Again, these factors,
and not necessarily light, may be acting as the primary cues
50
influencing brown trout behavior and activity.
There were large error bounds on the plots of local
activity. This is due in some cases to sample size, but
primarily to the nature of the activity pattern of these
fish. Because local activity was highly variable (long
periods of inactivity followed by short, rapid bursts of high
activity; or activity in a localized' area followed by
extended movements), even a large number of observations
would probably not tend to decrease the variance seen in this
data.
Mortality of brown trout following surgical implant of
transmitters was relatively high, as compared to mortality
rates reported in other studies. Schramm and Black (1984)
reported an average mortality rate to grass carp
Ctenopharyngodon idella following various surgical implant
procedures to be 31%, and Mulford (1984) Cited various
telemetric studies of striped bass Morone gaxatilie in which
mortality from tranmitter implant ranged from 30—46%. Both
Schramm and Black (1984) and Mulford (1984) indicated that
high water temperatures have an adverse effect on the
survival of fish following surgery.
In telemetry studies, two basic assumptions are made;
that transmitters do not adversely affect the behavior of
fish, and that the limited number of fish tracked ( due to
the increased expense of telemetry equipment ) adequately and
representatively reflect behaviors of the entire population
of fish under investigation. Some evidence that the first
assumption was met comes from capture of one fish at the end
51
of the study. The surgical incision had healed well, and the
fish was in excellent condition — able to elude an
electrofishing crew for almost one hour with strong upstream
and downstream swimming. Agreement between the findings of
this study and expectations for behavior of large brown trout
based on the current scientific literature provide some
evidence that the second assumption was met.
Management Implicationg
Ideally, research findings will positively influence
management decisions. Cobb (1933), uPOn determining that
brown trout displacement was primarily downstream, stressed
the; importance of stocking brown trout in a location where
they could run downstream into "good" water. Knowledge of
fish movements can also be of great importance in the
assessment of regulations and population estimates. One of
the major objectives of the present study was to relate the
movement of trophy brown trout in the South Branch of the Au
Sable River to a section of the river placed under catch—and—
release fishing regulations.
Over the course of this study, six of eight brown trout
tracked were observed to move out of the catch—and—release
section and into areas of the river under statewide trout
regulations. These fish were vulnerable to legal harvest by
anglers while outside the catch—and—release section.
However, four of five fish tracked during periods of heavy
summer fishing pressure (May—August) spent most of their
52
time inside the regulated area. Also, fish tracked during
fall and winter might have returned to this section of the
river during June and would again have been "protected".
This study shows that stream-resident trophy brown
trout can range over a section of river up to 34 km in
length, with the average home range being approximately 8 km
in size. Thus, special regulation areas (at least on the Au
Sable River) need to be at least 8 km long to ensure that the
home range of a single trophy fish is encompassed. When
choosing areas to be placed under special regulations, all of
the food and habitat requirements for trophy brown trout need
to be considered if the aim of the regulations is to produce
trophy fishing. Some areas might prove to be unsuitable for
trophy brown trout. In the long run though, no matter how
large an area we set aside, or how "good" an area it appears
to be, there will always be fish that will move out of the
area, or whose home ranges will straddle its boundaries.
Perhaps increases in the numbers of trophy fish harvested in
areas adjacent to these regulated areas could be counted as
an additional benefit of these quality fishing regulations.
53
Literature Cited
Alexander, G. R. 1977. Consumption of small trout by large
predatory brown trout in the North Branch of the Au
Sable River, Michigan. Michigan Department of Natural
Resources, Fisheries Research Report 1855, Ann Arbor.
Allen, K. R. 1951. The Horokiwi stream — a study of a trout
population. New Zealand Marine Department, Fisheries
Bulletin 10, Wellington.
Bachman, R. A. 1982. A growth model for drift-feeding
salmonids: a selective pressure for migration. Pages
128—135 i_ E. L. Brannon and E. O. Salo, editors.
Proceedings of the salmon and trout migratory behavior
symposium. University of Washington Press, Seattle.
Bachman, R. A. 1984. Foraging behavior of free—ranging wild
and hatchery brown trout in a stream. Transactions of
the American Fisheries Society 113:1-32.
Bosserman, W., and L. A. Higgins. 1969. Au Sable. Au Sable
River Watershed Study Council Report.
Chaston, I. 1969. Seasonal activity and feeding pattern of
brown trout (Salmo trutta) in a Dartmoor stream in
relation to availability of food. Journal of the
Fisheries Research Board of Canada 26:2165-2171.
54
Cobb, E. W. 1933. Results of trout tagging to determine
migrations and results from plants made. Transactions
of the American Fisheries Society 63:308—318.
Colwell, R. K. 1974. Predictability, constancy, and
contingency of periodic phenomena. Ecology 55:1148—
1153.
Coopes, G. F. 1974. AuSable River watershed project
biological report (1971—1973). Michigan Department of
Natural Resources, Fisheries Management Report 7,
Lansing.
Cunjak, R. A., and G. Power. 1986. Winter habitat
utilization by stream resident brook trout (Salvelinus
fontinalig) and brown trout (Salmo trutta). Canadian


Journal of Fisheries and Aquatic Sciences 43:1970-1981.
Diana, J. S., W. C. Mackay, and M. Ehrman. 1977. Movements
and habitat preference of northern pike (Esox lucius) in
Lac Ste. Anne, Alberta. Transactions of the American
Fisheries Society 106:560—565.
Diana, J. S. 1980. Diel activity pattern and swimming
speeds of northern pike (Esox lucius) in Lac Ste. Anne,
Alberta. Canadian Journal of Fisheries and Aquatic
Sciences 37:1454—1458.
55
Dill, L. M. 1978. An energy—based model of optimal feeding—
territory size. Theoretical Population Biology 14:396-
429.
Elliott, J. M. 1970. Diel changes in invertebrate drift and
the food of trout Salmo trutta L. Journal of Fish
Biology 2:161-165.
Elliott, J. M. 1975. Number of meals in a day, maximum
weight of food consumed in a day and maximum rate of
feeding for brown trout, Salmo trutta L. Freshwater
Biology 5:287—303.
Elliott, J. M. 1976. The energetics of feeding, metabolism
and growth of brown trout (Salmo trutta L.) in relation
to body weight, water temperature and ration size.
\
Journal of Animal Ecology 45:923—948.
Fausch, K. D., and R. J. White. 1981. Competition between
brook trout (Salvelinus fontinalis) and brown trout
(Salmo trutta) for positions in a Michigan stream.
Canadian Journal of Fisheries and Aquatic Sciences
38:1220-1227.
56
Gosse, J. C., and W. T. Helm. 1981. A method for measuring
microhabitat components for lotic fishes and its
application with regard to brown trout. Pages 138—149
i_ N. B. Armantrout, editor. Symposium on acquisition
and utilization of aquatic habitat inventory
information. American Fisheries Society Western
Division, Portland, Oregon.
Jenkins, T. M., Jr. 1969. Social structure, position choice,
and microdistribution of two trout species (Salmo trutta
and Salmo gairdneri) resident in mountain streams.
Animal Behaviour Monographs 2257—123.
Mesing, C. L., and A. M. Wicker. 1986. Home range,
spawning migrations, and homing of radio—tagged Florida
largemouth bass in two central Florida lakes.
Transactions of the American Fisheries Society 115:286—
295.
Mulford, C. J. 1984. Use of a surgical skin stapler to
quickly close incisions in striped bass. North American
Journal of Fisheries Management 4:571—573.
Oswald, R. L. 1978. The use of telemetry to study light
synchronization with feeding and gill ventilation rates
in Salmo trutta. Journal of Fish Biology 13:729—739.
57
Raleigh, R. F., and D. A. Duff. 1980. Trout stream habitat
improvement: ecology and hydrology. Pages 67-77 in W.
King, editor. Proceedings of Wild Trout II. Trout
Unlimited and Federation of Fly Fishermen, Denver,
Colorado.
Rankin, E. T. 1986. Habitat selection by smallmouth bass in
response to physical characteristics in a natural
stream. Transactions of the American Fisheries Society
115:322-334.
Schramm, H. L., and D. J. Black. 1984. Anesthesia and
surgical procedures for implanting radio transmitters
into grass carp. The Progressive Fish Culturist 46:185—
190.
Schuck, H. A. 1943. Survival, population density, growth,
and movement of the wild brown trout in Crystal Creek.
Transactions of the American Fisheries Society 73:209—
230.
Shetter, D. S. 1967. Observations on movements of wild trout
in two Michigan stream drainages. Michigan Department
of Natural Resources, Fisheries Research Report 1743,
Ann Arbor.
58
Shirvell, C. S., and R. G. Dungey. 1983. Microhabitats
chosen by brown trout for feeding and spawning in
rivers. Transactions of the American Fisheries Society
112:355-367.
Stalnaker, C. B., and J. L. Arnette, editors. 1976.
Methodologies for the determination of stream resource
flow requirements: an assessment. United States Fish
and Wildlife Service, Washington, District of Columbia.
Stauffer, T. E. 1977. A comparison of the diet and growth
of brown trout (§almg trutta) from the South Branch and
the Main Stream, Au Sable River, Michigan. Michigan
Department of Natural Resources, Fisheries Research
Report 1845, Ann Arbor.
Strauss, R. E. 1979. Reliability estimates for Ivlev's
electivity index, the forage ratio, and a proposed
linear index of food selection. Transactions of the
American Fisheries Society 108:344—352.
Swift, D. R. 1962. Activity cycles in the brown trout
(Salmo trutta L.): I. Fish feeding naturally.
Hydrobiologia 20:241—247.
White, R. J. 1975. In—stream management for wild trout.
Pages 48—57 in W. King, editor. Proceedings of Wild
Trout I. Trout Unlimited and Federation of Fly
Fishermen, Denver, Colorado.
59
Winter, J. D. 1983. Underwater biotelemetry. Pages 371-395
in L. A. Nielsen and D. L. Johnson, editors. Fisheries
Techniques. American Fisheries Society, Bethesda,
Maryland.
60
llillUllllllllllllllliillllllllllllll
92 7122