* *
* * tº
# * * * * * * : * ~ * -
sº, " … "
** * * * * *
a
E.
2 * ºr sP “sº ºr r *
* - a red
º
gº ºº we a º
* * * * * * * * * º - fºr sº-
is a tº
“º tº : & º *
****, *, *, - - - ºf pººr" 'º ºf “t ex ºf 1 * * * :
** * * *.*. - - - - - - 3. * - - 2 * * * * : * * * * * * tº a 'ſ - s.v. 3 v ºr a " a " " ' " ' "
º “º * * * - - - - - - - w --- --- - . . . "4: " : * . . . . . . "
- - - * f.º.ºr w. ** * * * * *...* * * - -
- - - - - - - G J - A * & * * * d - º - . . ) […] • *, * * * * *
- - - - * * * * - º Fº
* * ,
ge ºr rºw s ºw º ºf
* * B . tº wº º .*.*, *
‘. . .” - a rº, iſ ºw *** *****
* * * * * , ye " ' " " ºr w P :
* º -
º
- - º.
wº c t - * r * * * * * *- : ** ... * * * f. f. , .
sº tºº sº - - - - - ſº & J. - - - - a ºr ; - “s wº tº-º; ºn y.º.º.
- - - 5 - P * A { * * * * * *i. “ sº “," " - * ***
* . . m a 2
. . . .38 - sº * : * * * *
º º
it is, is, wº
tº anº D
- rº-
* * **** ****
s: ºn tº ºf
tº gº ºsº
wººl tº
- sº
º- {
fºr “sº- wº
lift ºr tº: * * *
* * * * * *
º, ºr sº º'
{ * , ºr " * : * r . -
. , , , is tº . ; ºf ºf
* \ g s a a * : s , : * º f tº º
- Nº." * : *ry tº ºn
ºf
* †
# ºr fººt tº
tº “tº sº. #' ººt ºº
º
º º f
º # * : *: º
* ,
sº
tº a tº f " ", , ºf º
...? Tº w sº 2 tº s' r" , , i.
t A & sº a sº ºft" ºr
* 3 & f * * **** -
- º



Assessment of the Channel Catfish Fishery
in Saginaw Bay, Lake Huron
by
Robert M. Lorantas
A thesis submitted in partial fulfillment of
the requirements for the degree of
Master of Science
in Fisheries
School of Natural Resources
The University of Michigan
1983
Thesis Committee:
Adjunct Professor William C. Latta, Chairman
Assistant Professor James S. Diana
Ex-officio examiner Richard D. Clark, Jr.
For
Justin, Nell
and
Kathy
i i
ACKNOWLEDGEMENTS
Financial support for this project was provided by
the Institute for Fisheries Research of the Michigan
Department of Natural Resources (MDNR) and the School of
Natural Resources at the University of Michigan. It was
challenging and enjoyable to become a part of a great
institution and great school. The abilities and expertise
of my committee members provided my greatest inspiration as
an aspiring professional. I would first like to thank my
chairman, Dr. W. Carl Latta, for providing me with the
opportunity and encouragement to engage in this study. I
would also like to thank Dr. James S. Diana for his interest
and encouragement. Thanks are extended to Mr. Richard
D. Clark, Jr., of the MDNR, for allowing me to utilize a
number of programs he developed, and for the generous
assistance and encouragement he provided. The critical
reviews and support of all committee members contributed
significantly to this study. My appreciation is extended to
Dr. Karl F. Lagler, his unfailing enthusiasm as my graduate
advisor made graduate work enjoyable. Indeed all of my
professors contributed to the progress of this research, and
I thank them. -
Numerous persons contributed to the development of
this project. Mr. Forrest Williams, Mr. Todd Williams,
Mr. Dennis Root, Mr. Richard Manor, and Mr. Samuel Morgan
supplied vessel time for sample collection. Mr. Ed Dootz
ii i
provided laboratory space in the Dental Materials Department
at the University of Michigan. Mr. Andrew Nuhfer, Mr. Don
Nelson, Mr. Gale Jamson, and Mr. John Weber all of the MDNR
provided commercial and sportfishing catch data along with
useful information concerning the fisheries of Saginaw Bay.
Mr. Paul Haack, of the Great Lakes Fishery Laboratory of the
U. S. Fish and Wildlife Service, provided historical
commercial catch records. Fellow student, Tim Hill,
assisted in data collection. Many not mentioned here also
contributed. I thank everyone for their interest, kindness,
and time. Lastly, I extend my heart felt thanks to my
parents and sister for their special support and
encouragement which allowed me to pursue my life long
interests.
i v
ABSTRACT
channel catfish Ictalurus punctatus ranked second in
weight harvested by commercial fishermen (231,000 kg), and
third in number caught by sport fishermen (60,000) in
Saginaw Bay in 1981. The commercial fishery employs trap
nets, seines, and set hooks. Mean annual catch per unit
effort for all gear types has increased in the past decade,
in comparison to prior decades, and has lead commercial
fishermen to request licensing of additional gear. The
commercial fishery was assessed using a dynamic pool model,
and an extension of the model was used to investigate the
dynamics of gear competition. Growth and total mortality
parameters, estimated from four management areas, were
pooled for model analysis since no significant differences
in these vital statistics were detected after the age of
complete recruitment to the fishery. Parameters of the
von Bertalanffy growth equation were estimated using mean
back-calculated lengths at age derived from fin spine
sections. Total instantaneous mortality was estimated from
the slope of the descending limb of a catch C Ul Iſ Ve . Fishing
mortalities for each commercial gear type and for
sportfishing gear were estimated by partitioning the total
fishing mortality in proportion to the catch from that gear.
Pooling all areas yielded a von Bertalanffy equation of the
form Px --> 921 (1-e (-0.09(x-0.35)))
, and a total instantaneous
mortality of 0.67. Model predictions indicated that yield
to the commercial fishery and to the sport fishery could be
increased by increasing the minimum commercial size limit
and/or reducing the commercial fishing mortality.
Simulations also indicated that an increase in fishing
mortality by any one gear type increased yield to that gear
type, but reduced yield to all other gear types. The
tenuous nature of the estimates of sportfishing mortality
and natural mortality preclude specific management
recommendations.
vi
TABLE OF CONTENTS
DEDICATION . . . . . . . . . . .
ACKNOWLEDGEMENTS . . . . . . . .
AB STRACT O © O O O O O O O Q Q O
LI ST OF TABLES O O O O Q © © C ©
LIST OF FIGUREs . . . . . . . .
LIST OF APPENDICES . . . . . . .
I NTRODUCTI ON O © O © O O O O Q O
BACKGROUND O O O O O © Q Q O © O
METHODS
RESULTS
Data Collection . . . . .
Fin Spine Preparation . .
Back-calculation Using Fin Spines .
Age Composition and Size
Catch . . . . . . . . . .
Growth in Length . . . .
Weight-Length Relation .
Mortality © © © G © O O ©
Yield Analysis . . . . .
Age Composition and Size
Catch © O O O º O O e © O
Growth in Length . . . .
Weight-Length Relation .
Mortality . . . . . . . . .
Yield . . . . . . . . . .
DISCUSS
ION O O © O O º O © © O O
Age Composition and Size
Catch O O © O © Q O © © ©
Structure
Q
Structure
Structure
of
i i
ii i
23
25
27
27
29
44
44
vi i
Growth in Length . . © O
weight-Length Relation .
Total Mortality Rate e e
Fishing Mortality Rates .
yield . • e e o e e s e e
LITERATURE CITED . . . . . . . .
45
49
49
51
54
60
viii
Table
LIST OF TABLES
Mean annual catch per unit effort for trap
nets, seines, and set hooks and mean annual
yield for all commercial gear types combined
during selected time periods. . . . . . . . .
Sampling dates, number of lifts sampled,
average trap net pot height, pot stretch mesh
size, number of specimens sampled for total
length measurement, and number of specimens
subsampled for weight, spine removal, and sex
from each grid in 1981 . . . . . . . . . . . .
Mean back-calculated lengths-at-age derived
from the the posterior and anterior fields of
spine sections from the distal end of the
basal recess, mean empirical lengths-at-age,
and number of specimens aged in grid 1608. . .
Annual mail survey estimates of sport harvest
of ic talurids (in number), fraction of channel
catfish in the commercial harvest ( in number),
and adjusted estimate of yield (in weight) of
sport-harvested channel catfish in Saginaw Bay
(1975-1981) . . . . . . . . . . . . . . . . . .
Estimated age distributions of channel catfish
caught in comercial trap nets for each grid
sampled and for all grids pooled. . . . . . .
Yield to each gear type, fraction of total
yield attributable to each gear type, and
estimate of the instantaneous fishing
mortalities due to each gear type in 1981.
Range of FS + M values used to examine the
sensitivity of yield per 1000 recruits as a
function of FC, and range of FS + M (and FC)
values used to examine the sensitivity of
yield per 1000 recruits as a function of *c.
Mean back-calculated lengths at age of channel
Page
10
12
19
23
31
39
catfish from Saginaw Bay in 1981 and 1971, and
from other selected locations. . . . . . . .
48
ix
9.
Instantaneous total mortality rates estimated
for Saginaw Bay in 1981 and 1971, and for
other selected locations. . . . . . . . . . .
52
10.
1 1.
LIST OF FIGURES
Map of Saginaw Bay, Lake Huron, illustrating
the grid system used in sampling. . . . . . .
Annual catch, effort and catch per unit effort
for trap nets, seines, and set hooks from 1929
to 1981 . O O O O Q O O O O O O O O º © O O Q O
Length frequency polygon of channel catfish
caught in commercial trap nets for all grids
pooled O O O O O O O O © Q © © © C O O O Q O ©
Estimated von Bertalanffy growth curve and
empirical mean back-calculated lengths at age
with 95% confidence limits. . . . . . . . . .
Catch curve with regression line fitted to the
descending limb (grids combined). . . . . . .
Yield isopleths per 1000 recruits as a
function of FC and *c with Ma 0. 1 and FS= 0.06.
Yield per 1000 recruits as a function of FC
with FS= 0.06, M= 0. 1, and l. = 38.1 mm; and for a
range of Fs - M values, with i-381 mm. . . .
Yield per 1000 recruits to sport gear, seines,
set hooks, and trap nets as a function of FR,
where FS= 0.06, FE=0.05, FH=0. 14, M=0. 1, and
le-38 Iſlm • . . . . . . . . . . . . . . . . . .
Yield per 1000 recruits to sport gear, seines,
set hooks, and trap nets as a function of FH
where FS= 0.06, FE=0.05, FR=0. 32, M=0. 1, and
le-38 Imm - - - - - - - - - - - - - - - - - - -
Yield per 1000 recruits to sport gear, seines,
set hooks, and trap nets as a function of FE
where FS= 0.06, FH=0. 14, FR= 0.32, M=0. 1, and
le-38 Iſlul - - - - - - - - - - - - - - - - - - -
Yield per 1000 recruits to the commercial
fishery and sport fishery as a function of le
where FS= 0.06, FC= 0.51, and M=0. 1 . . . . . . .
Page
24
28
30
32
34
36
37
38
4 1
xi
12. Yield per 1000 recruits as a function
of l
with FC=0. 51, FS= 0.06, and M-0. 1; and for
range of FS + M values, with FC=0.51.
g
42
xii
LIST OF APPENDI CES
Appendix
A.
Mean back calculated lengths at age for both
sexes in each grid and for sexes combined.
(Sample size and 95% confidence interval are
indicated for each length.) . . . . . . . . .
Polynomials fitted to back-calculated lengths
at all ages for each sex in each grid and for
sexes pooled within grids. (Sample sizes and
coefficients of determination are indicated. )
Linearized weight-length regressions for each
sex in each grid, for sexes pooled within
grids, and for sexes and grids pooled. (Sample
sizes and coefficients of determination are
indicated.) . . . . . . . . . . . . . . .
Equations fitted to the descending limbs of
catch curves for each grid and grids combined.
(Sample sizes and coefficients of
determination are indicated. ) . . . . .
Page
xiii
INTRODUCTION
Channel catfish Ictalurus punctatus in Saginaw Bay,
Lake Huron, Support both commercial fishing and
sport fishing. In 1981, sport fishermen harvested a ſl
estimated 29,000 kg and commercial fishermen a reported
231,000 kg. Channel catfish are a valuable segment of the
commercial fishery and ranked second in weight harvested in
1981; common carp Cyprinus carpio ranked first (314,000 kg),
and yellow perch Perca flavescens ranked third (80,000 kg).
Yellow perch ranked first in number of fish harvested by
sport_fishermen, sunfish species ranked second (these
included mostly bluegills Lepomis machrochi rus and
pumpkinseeds L. gibbosus), and channel catfish ranked third.
It appears that there exists sufficient social and economic
impetus for the continued coexistañce of recreational and
commercial fisheries throughout the Great Lakes. However,
allocation of fishery resources to their respective users
has been a difficult problem to solve (Franc is et al. 1979;
Bishop and Samples 1980; Talhelm 1979). In the existing
fishing regime, the assessment of changes in yield or effort
levels of sport and commercial fisheries becomes even more
crucial to resource managers if yields acceptable to both
users are to be maintained.
Efforts to quantitatively evaluate multiple use of a
common fishery resource are limited (Clark and Huang 1983;
Low 1982). Yet, managing only the commercial fishery
without regard to the sport fishery or vice versa may be
futile in achieving a desired management objective.
Although sport fishing harvest records provide only an index
of the harvest of channel catfish in Saginaw Bay, available
records were utilized in this assessment. The objectives of
this study were: (1) to examine the history of sport and
commercial fisheries along with their regulation, (2) to
collect timely information concerning growth and mortality
rates of channel catfish in different areas of Saginaw Bay,
(3) to determine whether these vital statistics differ
between areas, and (4) to evaluate some management
strategies using available in formation in a
yield-per-recruit model.
BACKGROUND
Beeton et al. (1967) described the location,
morphometry, and limnological aspects of Saginaw Bay. Hile
(1959) described the multi-species and multi-gear character
of the fisheries in Saginaw Bay and documented changes in
species composition and gear composition utilizing
commercial catch records. The Great Lakes Basin Commission
(1975) summarized historical changes in sport and commercial
fisheries as well as associated changes in water chemistry
and aquatic biota and provided a broad perspective of the
dynamic nature of the Bay. Characteristics and regulation
of the commercial and sport fisheries for channel catfish
make them unique among fisheries of the Great Lakes. . The
commercial fishery employs three major gear types: trap
nets, seines, and set hooks. Trap nets harvest all
commercially available species in the Bay, seines primarily
harvest common carp and channel catfish, and set hooks
harvest channel catfish exclusively. The contribution of
channel catfish in weight to the commercial harvest in 1981
was 63% for trap nets, 9% for se ines, and 28% for set hooks.
All units of gear are licensed by the state; the type and
amount of gear licensed have evolved through restrictions
imposed by the state and by fishermen utilization of the
various gear types. In 1981, 400 trap nets, 16 seines, and
39, 400 set hooks were licensed to 28 commercial operators.
Additional units of gear have not been available for
licensing since 1970 in areas of the Bay open to commercial
fishing (Great Lakes Fishery Commission 1971 a ).
Commercially fishable areas of the Bay have been largely
confined to the Inner Bay, that area southwest of a line
connecting Sand Point and Point Lookout (Fig. 1). Minimum
size and weight restrictions relating to commercial harvest
of channel catfish date from at least 1895 during which a
0.45 kg ( 1.0 lb) minimum weight limit was in effect. This
limit evolved into a 0.91 kg (2.0 lb) minimum weight limit
in 1929, a 432 mm (17 inch) minimum size limit in 1945, and
a 381 mm (15 inch) minimum size limit in 1960. The latter
regulation remains in effect. There a e In O Sea SOIn
restrictions governing commercial harvest.
Records of the annual commercial harvest of channel
catfish date from 1919 (Baldwin et al. 1979). Since 1929,
trap nets have accounted for the bulk of the commercial
catch, except during the period 1959 - 1965 when set hooks
produced the largest yields. Seines ranked second in weight
harvested until about 1956 and have since accounted for the
lowest yields. Change in effort was the apparent cause of
fluctuation in yield to each gear until the mid 1960's
(Fig. 2, A and B). Historically, the fishery has been
uncharacteristically stable among the commercially important
fisheries of the Bay (Hile 1959). Although catch, effort,
and catch per unit effort (CPUE) fluctuated from very low to
very high levels between 1940 and 1970, Eshen roder and Haas
(1974) concluded that these changes were not indicative of
significant changes or trends in population abundance
Towgs City
!
point Lookours
6 O 6 isost
º SAGINAW BAY
- LAKE HURON
Figure 1. Map of Saginaw Bay, Lake Huron, illustrating the
grid system used in sampling.

:
200-
A
El CºA
T R A P Nº T3
Aſ
A
$0
| IAP Mºrº llºrs ºf 1,000 - B
sºng HAuks. 199 - - - - - - de
sky Mook.utºrs. GMEcººp. 4.129 ... ',
20- • ".
10-
4 sº –’ Q9 \ ..?- o *
{ ams sº Z N, T • * * \,” º © \ ^ se
{ \ , *e ... • * * e = e s e - " " s .* e e ^
•. - - - " * @ * ~ * * * = /
() -——---------------—I----—-——--I--——-——- I----—— I———- I---
1920 19 JO 1949 1999 1900 1979
YEAR
40 -wr-
| | IMAP Mºrs (to/LIFT) C
Figure 2. Annual catch, effort and catch per unit effort
for trap nets, seines, and set hooks from 1929 to
1981. (See Hile (1962) for specific descriptions
of effort units. )




(Fig. 2). The Great Lakes Basin Commission (1975), however,
postulated that the downward trend in catch during the mid
1960's was probably due to a decline in abundance, since at
this time the value per pound was increasing. The decline
in catch, effort, and CPUE during the mid 1960's may,
likewise be associated with fear of mercury contamination
evident in catfish from other areas of the Great Lakes
during that era (Great Lakes Fishery Commission 1971b). The
mean annual catch in the period from 1970 to 1981 has
increased; this increase has been associated with an
increase in the mean annual CPUE for all gear types (Table
1). Thus, in contrast to the previous decade, these changes
are probably indicative of an increase in abundance. The
increases in CPUE have prompted requests by commercial
fishermen for the licensing of additional gear (John Weber,
Michigan Department of Natural Resources, personal
communication). A goal of this assessment was to determine
Table 1. Mean annual catch per unit effort (CPUE) for
trap nets, seines, and set hooks and mean
annual yield for all commercial gear types
combined during selected time periods.
Mean annual CPUE
Mean annual
Year's Trap net Seine Set hook yield (kg)
1940-49 12 1.15 86 133, 655
1950-59 15 123 90 126,795
1960-69 13 56 64 75,261
1970–81 42 130 142 166, 153
whether expenditure of aditional commercial effort, perhaps
through the licensing of additional commercial gear, was an
appropriate management alternative.
Sportfishing has been found to have significant
impacts on fisheries exploited both commercially and by
sportfishermen (McHugh 1980). Sport fishing demand for
channel catfish has not been assessed in Michigan. However,
since 1975, estimates of the sport harvest of catfish of all
species have been made by the Michigan Department of Natural
Resources. From 1975 to 1981 no distinct trends were
evident in sport fishery yield of channel catfish from
Saginaw Bay. (More specific information concerning the
estimation of sport fishery yield of channel catfish will be
provided in later sections.) The sport fishery for channel
catfish has no season, size, or possession restrictions; and
there are no limitations on the number of sportfishing
licenses sold. In the present study, the impact of
sportfishing mortality and variation in sport fishing
mortality on channel catfish was considered in assessing the
commercial fishery.
METHODS
Data collection
To estimate parameters which describe the growth and
mortality of channel catfish, commercial trap net catches
were sampled in each of four management areas (in grids
1608, 1509, 1507, and 1606) within the Inner Bay (Fig. 1).
The total length of all channel catfish from a trap net
catch was recorded to the nearest millimeter. A subsample
of approximately 10 to 20 specimens per 25 mm group were
collected for further analysis. These fish were weighed to
the nearest gram using a spring scale, dissected t O
determine sex, and the left pectoral spine was removed for
growth analysis. Sampling dates, the number of lifts
sampled, characteristics of the commercial sampling gear,
number of specimens collected, and number of specimens
subsampled from each grid are listed in Table 2.
Ein Spine Preparation
Specimens were aged from annuli observed on sections
of pectoral fin spines (Sneed 1951). Use of pectoral fin
spine sections for age determination of channel catfish has
been validated with known age specimens by Sneed (1951),
Marzolf (1955), and Prentice and Whiteside (1974). Spines
were removed as described by Sneed (1951), air dried, and
sectioned at the distal end of the basal recess. The
thickness of a section was determined by the thickness of a
spacer between two cutting discs similar to the method of
10
Table 2. Sampling dates, number of lifts sampled, average
trap net pot height, pot stretch mesh size,
number of specimens sampled for total length
measurement, and number of specimens subsampled
for weight, spine removal, and sex from each
grid in 1981. When a pot consisted of two
mesh sizes the size comprising the lesser
area of the pot is listed in parenthesis.
Number
Number Mean POt Number of
of pot mesh of specimens
lifts height size lengths sub-
Grid Month/day sampled (m) (mm) sampled sampled
1608 5/16-6/13 2 3.05 6.35 (4.44) 1051 264
1509 6/2-6/20 4 3.05 6. 35 (4.44) 740 21 1
1507 7/16-8/4 18 1 - 07 4.76 (5. 17) 640 206
1606 8/1 6 2. 13 4. 76 542 236
Chugunova (1959). I used aluminum oxide cutting discs and a
0.52 mm spacer rotated at 2750 RPM with a dental lathe. A
stream of water directed on the discs served as a lubricant
and coolant. After sawing, sections were air dried, and
then mounted on cellulose acetate slides with viscous
cyanoacrylate adhesive. Spine sections prepared in this
manner could be viewed with transmitted light using a scale
projector. Sections too thick for light transmission could
be made thinner with a fine toothed file.
Back-calculation. Using Ein Spines
Back-calculated lengths were computed using the
formula described by Everhart and Youngs (1981). Several
problems have been encountered in calculating lengths using
1 1
spine sections. Muncy (1959) and Marzolf (1955) identified
the following causes of bias:
(1) Pectoral spines are tapered, and sections obtained
at the basal recess, which expands distally with
increasing age, decrease in relative size with
increasing age.
(2) The maximum expanded portion of annuli in the
commonly measured posterior field does not always
lie in a straight line along the maximum radius.
(3) The center of the lumen of the spine, commonly used
as the origin in making radial measurements to
annuli, does not always correspond to the center of
the first annulus.
(4) Spine sections are not always sectioned precisely
perpendicularly.
DeRoth (1965) addressed the first problem by
section ing spines at the same relative location. However, I
had difficulty in consistently determining the appropriate
point of sectioning. DeRoth's method also produced sections
without a lumen or complete first annulus, making radial
measurement difficult. To counter both the first and the
second problems, Marzolf (1955) recommended measurement of
the anterior radius, which he suggested was less affected by
spine taper. To check this assumption I computed intercepts
(correction factors) and mean back-calculated lengths at age
for specimens from grid 1608 for both posterior and anterior
fields of spines sectioned at the distal end of the basal
recess. Total length was regressed on each spine radius; a
straight line adequately described these relationships. The
intercepts estimated for the anterior and posterior fields
were -46. 1 and - 198.7 respectively. The anterior radius
12
Table 3. Mean back-calculated lengths (mm) at age derived
from the posterior and anterior fields of spine
sections from the basal recess, mean empirical
lengths (mm) at age, and number of specimens
aged in grid 1608.
1 2 3 4. 5 6 7 8 9 1 0 1 1 12
Calculated .
posterior - 100 1 1 140 237 300 350 410 460 499 542 580 597
Calculated -
anterior 51 129 196 254 .307 353 4 13 460 500 542 582 597
Empirical ... 144 228 255 316 340 398 479 505 539 589 597
Number of
specimens O 1 2 28 1 46 9 3 12 24 9 1
produced calculated lengths similar to empirical lengths at
age, while the posterior radius produced calculated lengths
which were negative and erroneous (Table 3).
The degree to which the third and fourth problems
affect the analysis depends upon how care fully spines are
measured and sectioned. I took the following precautions to
mitigate or eliminate all the previously described problems:
(1) The anterior radius was measured which alleviated
problems associated with spine taper at the basal
Iſ eCe SS • .
(2) Annuli in the anterior field were concentric with
respect to the center of the first annulus of the
section. Therefore, measurement along this field
reduced the error associated with measurement of
expanded annuli in the posterior field.
(3) The approximate center of the first annulus was
used as the origin for radial measurement which
decreased error associated with U.S e of a ſl
irregularly positioned or irregularly shaped lumen.
13
(4) Spines were marked prior to sectioning which
facilitated cutting perpendicular sections and
reduced mechanical difficulties associated with
positioning the spine during the section ing
process. -
Age Composition and Size Structure of the Catch
Age distributions of catches were estimated using
age-length keys (Allen 1966). Age-length keys were
constructed for each grid sampled and applied to the length
distribution of the catch from that grid. This procedure
afforded independent estimation of the age distribution of
the catch from each grid, and thus avoided possible bias due
to age composition differences among grids (Kimmura 1977).
Estimated age distributions from each grid were then
compared with a G-test (Sokal and Rohlf 1981). The computer
program, CHITAB, facilitated computation of the attained
significance level for the G-test (Statistical Research
Laboratory 1976). A length-frequency polygon of the
combined catch of all grids was also constructed t O
determine the size of complete recruitment to the trap net
fishery.
Growth in Length
Total length was regressed on anterior spine radius
for each sex in each grid using least squares. A straight
line adequately described these relationships. The
intercepts of the regression equations were compared between
sexes and grids. Data were pooled for grids and sexes with
14
similar intercepts and a common intercept (correction
factor) was used to compute back-calculated lengths. Back-
calculated lengths were used to check for growth differences
which would preclude use of an average growth function to
describe growth of each sex and growth in each of the four
management areas of the Bay. Although the von Bertalanffy
function is appropriate for and was fitted to the
length-at-age data in this study, interpretations of the
parameters of this function for comparative purposes and
statistical tests concerning these parameters have been
controversial (Gallucci and Quinn 1979; Kingsley et
al. 1980). To statistically compare growth, back-calculated
lengths were regressed on age for both sexes in each grid.
Visual inspection of the relationship between back-
calculated length and age, and residual analysis indicated
that a second degree polynomial adequately described this
relationship. !
2
y - Bo B12 + B2x",
1 2
where
Y = back-calculated length in mm;
X = age in years;
Bo = intercept;
B = linear effect coefficient;
B2 = curvature effect coefficient.
Regression equations and regression parameters were compared
to test for growth differences. Linear and curvature effect
coefficients, which describe the shape of the length-age
15
relations, were compared between regression equations to
check for differences in growth rate. If differences did
not exist, intercepts were compared to check for differences
in the magnitude of lengths-at-age.
weight-Length Relation
The weight-length relation, w = al”, was linear ized
with a natural log transformation:
loge (w) - ? loge (a) * beloge (l),
where
w = weight in kg;
l = length in mm;
loge (a) = intercept;
b = slope.
The natural log of weight was regressed on the natural log
of length for each sex in each grid, and the equality of
these regression equations was tested. All weight-length
data were later pooled to estimate an average weight-length
relation for use in yield calculations.
Mortality
The total instantaneous mortality rate (Z) was
estimated for each grid sampled by determining the slope of
a straight line fitted to the descending limb of a catch
curve (Ricker 1975). Everhart and Youngs (1981) recommended
fitting the regression line to age groups which include the
age group one year older than the modal age group and extend
to the age group one year younger than the oldest age group
16
captured. These recommendations were intended to reduce
bias due to gear selectivity and small sample variation.
Catch curves in this study, however, exhibited considerable
variability. Variability near the domes did not permit
accurate determination of the first fully recruited age
gr Oup, therefore in fitting regression lines to the
descending limbs, the modal age group was assumed to be
fully recruited. The oldest age group was also assumed to
be appropriately represented and was used in calculating the
slope of the descending limb. Instantaneous total mortality
rates were compared by testing for differences in the slope
parameters of the regression equations.
All regression parameters, test statistics, and
attained significance levels were computed using MIDAS, a
computer program for data analysis developed by the
Statistical Research Laboratory at the University of
Michigan (Fox and Guire 1976). All statistical tests were
based upon a 5% level of significance.
The total instantaneous fishing mortality rate (FT)
was estimated by subtracting the instantaneous natural
mortality rate (M) from the instantaneous total mortality
rate (Z):
FT = Z - M.
I could not estimate the natural mortality rate from my
data, so I used the rate of 0.1 reported by Eshen roder and
Haas (1974). The instantaneous fishing mortality for each
gear type or fishery was estimated by multiplying the total
17
instantaneous fishing mortality by the proportion of yield
attributable to each gear or fishery for 1981. Partitioning
forces of fishing mortality in this manner assumes that
these competing risks of capture are independent, and thus
additive. Therefore,
FT = FS + FR + FE + FH,
where
FT = instantaneous total fishing mortality;
FS = instantaneous sport fishing mortality;
FR = instantaneous trap net fishing mortality;
FE = instantaneous seine fishing mortality;
FH = instantaneous set hook fishing mortality.
The Michigan Department of Natural Resources annually
compiles the weight of fish harvested by each commercial
gear type from reports submitted by commercial fishermen,
and annually conducts a mail survey to estimate numbers of
fish harvested by sport fishermen. Several adjustments of
the mail survey data had to be made to make this information
suitable for my use. First, the survey estimated the
combined harvest of all ic talurids' from Saginaw Bay. To
estimate the sport harvest of channel catfish alone, I
assumed the fraction of channel catfish in the combined
commercial harvest was the same as in the sport harvest.
(The weight harvested commercially was available to the
'Ictalurus spp. in Saginaw Bay includes nebulosus,
the brown bullhead; natalis, the yellow bullhead; and
punctatus, the channel catfish.
18
species level. ) Second, sport harvest estimates were
reported in numbers of fish harvested, while commercial
harvest was reported in weight. Thus, I had to estimate the
average weights of fish harvested in the respective
fisheries to get harvest in common units of measure. Using
pooled weight and length data from this study I calculated
the average weight of a channel catfish in the commercial
catch to be 1.0 kg, 2.2 lb (fish 2 381 mm, 15 inches in
length), and the average weight of a sport-harvested fish to
be 0.4 kg (0.9 lb) (assumed to be the average weight of fish
2 299 mm, 12 inches, in length). The average weight of a
bullhead species was assumed to be 0.2 kg (0.4 lb), as
reported for brown bullheads by Blumer (1982). Finally,
Rybicki and Keller (1978) reported that mail survey
estimates were inflated for some species by factors of 5 to
20. Overestimation of harvest by a factor of 5 has been
substantiated by other research and is used as a standard
correction for mail survey estimates of catch for most
species in Michigan (Talhelm et al. 1979). Thus, as a
final adjustment, I reduced estimates of sport harvest in
weight of channel catfish to one-fifth of their estimated
values. Putting all these adjustments together it was
possible to estimate the weight of channel catfish harvested
by sport fishermen (Table 4). The 1981 sport harvest
estimate and reported commercial harvest by gear type were
then used to partition FT.
19
Table 4. Annual mail survey estimates of sport harvest of
ic talurids, fraction of channel catfish in the
commercial harvest, and adjusted estimate of
sport-harvested channel catfish in Saginaw Bay.
Mail survey Fraction of Adjusted yield
estimate of channel catfish in of sport-harvested
ictalurids commercial harvest channel catfish
Year (numbers) (numbers) (kg)
1975 523,430 0. 587 29, 919
1976 426,880 0. 715 29, 721
1977 263,235 0. 794 20, 353
1978 297, 120 - 0.94 1 27, 225
1979 337,600 0.964 31, 69 l
1980 205,700 0. 982 19, 670
1981 3 10,500 0. 975 29, 479
20
Yield Analysis
An idealized view of the dynamics of a catfish
fishery could be described schematically as follows:

Natural Mortality
Trap net Mortality
Set hook Mortality
Seine Mortality
Sport Fishing Mortality
Recruitment —- Population
Growth —- Biomass
E
Additions to the population come in the form of growth and
recruitment and losses arise from natural, trap net, seine,
set hook, and sport_fishing mortality. This approach is an
extension of the cohort-yield approach of Beverton and Holt
(1957) and Ricker (1975), and accounts for fishing mortality
from four independent sources rather than one. The model I
used, developed by R. D. Clark, Jr. of the Michigan
Department of Natural Resources, was parameterized so that
it accommodated four independent SOU iſ C eS of fishing
mortality. The model was used to calculate the yield per
number of fish recruited (N) at the age of first
vulnerability to fishing (x,). (In all analyses the number
of fish recruited at age x - was set at 1000.) From age *r
r
to the age at which fish were harvested (xe), losses were
assigned to natural mortality. When fish reached an age (x)
greater than *c losses were assigned both to natural
mortality and fishing mortality due to all gear types (FT):
21
dN/dx
-MN xrsks.Xe
— (FT-FM) N x^xe
dN/dx
Catch rate was then described in terms of each gear type
dca/dx E -Fºn x^xe .
Integrating and combining these equations resulted in a
catch equation where the number of fish at each age caught
by each gear (°x, g) Wa S :
Z
C -N, (Faſz, ) (1-et x);
X , 9
where
2x = (FT-M) [(x+1)-x ).
The von Bertalanffy equation of growth in length was used to
calculate mean length at age and the weight-length relation
was used to estimate the mean weight at age. Mean weights
at age multiplied by the mean number harvested at age for
each gear produced the weight harvested at each age by each
gear. Summing the weights harvested at all ages for each
individual gear type produced the yield in weight by gear
per 1000 recruits.
The model utilized length at capture (le) and length
at recruitment (lr) data which were converted to age at
capture (xe) and age at recruitment (xr). The length at
recruitment (lr) to all gear types was defined to be the
mean length of fish less than or equal to the modal length
of fish in the length frequency distribution of the
commercial catch. This length was also assumed to be the
22
size acceptable to sport_fishermen. A more detailed
mathematical description of a similar model is presented in
Clark and Huang (1983).
Yield to the commercial fishery per 1000 recruits was
sensitive to changes in FS and M. These parameters were
combined (FS + M) to examine this sensitivity. Yield to the
commercial fishery per 1000 recruits was examined for FS + M
values of: 0. 1, 0.16, 0.2, 0.25, 0.3, 0.35, 0.4, and 0.45.
Yield-per-recruit analyses were performed in conjunction
with various commercial minimum size limit regimes and
commercial fishing mortality regimes to identify conditions
under which maximum yield occurred. Only changes in
regulation of the commercial fishery were considered,
however the implications of these changes were evaluated for
both sport and commercial fisheries.
RESULTS
Age Composition and Size Structure of the Catch
i Estimated age distributions of channel catfish caught
in commercial trap nets were significantly different between
grids (G-test, P-0.01). The modal age of fish captured in
grids 1509 and 1608 was age 6, in grid 1507 age 7, and in
grid 1606 age 5. In grids 1608 and 1606 a greater
proportion of individuals less than age 6 were captured than
in grids 1509 and 1507 (Table 5). Differences in catch
characteristics were probably associated with differences in
availability of channel catfish at the time of sampling, and
differences in the selective properties of the sampling
gear. With all ages pooled, age 6 was the modal age
captured (table 5), and 340 mm (13 inches) was the modal
length of channel catfish captured (Fig. 3).
Table 5. Estimated age distributions of channel catfish
caught in commercial trap nets for each grid
sampled and for all grids pooled.
Age
Grid 1 2 3 4 5 6 7 8 9 1 0 1 1 12 13
1608 O O. 1 0.6 8.6 2.5 69.0 4.7 1.6 4.2 7.0 1.2 0.5 0
1509 0 0 0 5. O 0.7 60. 4 12. 7 2. 4 5 - 9 8. 7 2. 7 1. 4 0 . 1
1507 0 0 0 1.0 9.5 12.7 56.4 9.5 1. 1 4.6 3.6 1.3 0.3
1606 0 0.6 5.7 17.5 38.7 4.6 22.0 3.3 1.3 3.0 2.2 1. 0
All 0 0 . 1 1. 2 7.7 10. 2 4 3. 1 20.9 3.8 3.4 6. 2 2. 3 1.0 0 . 1
23
24
0.40
0. 55-
0.30-
0.25-
0.20-
0, 15-
0.05 -
0,00- T lſ Uſ N I T *-*
100 130 260 $40 420, 500 580 660 740 820
LENGTH INTERVAL MIDPOINT (MM)
Figure 3. Length frequency polygon of channel catfish
caught in commercial trap nets for all grids
pooled (interval width = 40.0 mm).

25
Growth in Length
Length-age regressions were significantly different
between sexes and grids (f-test, P-0.0001). To identify
where these differences occurred, regressions between sexes
within grids were compared. these regressions were not
significantly different in grids 1507 (f-test, P=0. 82) and
1606 (f-test, P=0.57), but were significantly different in
grids 1509 (f-test, P-0.0001) and 1608 (f-test, P=0.0003).
There were significant differences between shape parameters
(B1's and B2's) in grids 1509 (f-test, P-0.0001) and 1608
(f-test, P=0.0001) which indicated that differences in
growth rate existed between males and females in these
grids. Mean back-calculated lengths at age were greater for
females after age 6 in grid 1608 and greater for males at
all ages in grid 1509 (Appendix A). These conflicting
growth differences were probably due to differential
vulnerability of the sexes to capture during the spawning
Sea SO In rather than real growth differences. Although
statistically significant differences in growth rate existed
between sexes in grids 1509 and 1608, sexes were pooled to
check for spatial growth differences.
With S eX eS pooled length-age regressions were
significantly different between grids (f-test, P-0.0001).
The shape parameters of the regression equations were not
significantly different (f-test, P=0.06), however the
intercepts (Bo's) were significantly different (f-test,
P<0.0001). Pairwise comparisons indicated that regression
26
equations were identical in grids 1507 and 1509 (f-test,
P=0 . 89). The intercepts of the length-age regressions in
grids 1608 and 1606 were less than the intercepts in grids
1509 and 1507 (Appendix B). In grid 1608, mean back-
calculated lengths at age were less from age 1 to age 9 than
in other grids (Appendix A). These differences were
probably associated with differences in vulnerability of
various size classes to capture during the spawning season.
In grid 1606, mean back-calculated lengths at age were less
than those in grids 1507 and 1509 (Appendix A ). These
differences were probably associated with differences in
availability of various size classes, as well as differences
in selective properties of the sampling gear. A regression
equation fitted to back-calculated lengths versus age after
age 6, the age of complete recruitment to the fishing gear,
indicated that the length-age relation was the same between
all grids (f-test, P=0.3). Differences in length-at-age did
not appear great enough to treat growth between sexes or
management areas separately, therefore growth was described
by a single function in the yield-per-recruit analysis. The
von Bertalanffy function was fitted to the mean back-
calculated lengths at age, for grids and sexes pooled, using
the method of Rafail (1973). The equation derived was:
Lx * 921 (1-e (-0.09(x-0.35))),
27
where
L
=
length at age x (mm);
age in years.
The model agreed quite closely with the empirical data
(Fig. 4).
Weight-Length Relation
Weight-length regressions between sexes and grids
were significantly different (f-test, P30.0001). To
identify where these differences occurred, regressions
between sexes within grids were compared. Regressions
between sexes were not significantly different [1608 (f-
test, P=0.34), 1509 (f-test, P=0. 10), 1507 (f-test, P=0. 10),
and 1606 (f-test, P=0. 10) ). However, with sexes pooled
equations were significantly different between grids (f-
test, P-0.0001). Differences in weight-length equations
probably reflect differences in condition associated with
the time of year the samples were collected (Table 2).
Pooling all grids and sexes yielded the following weight-
length relation used to compute mean weight from mean length
in yield calculations:
loge (w) = -20. 71 + 3.36°loge (l).
Regression equations for each sex within each grid and for
sexes pooled within grids are listed in Appendix C.
Mortality
Total instantaneous mortality rate S and 95%
28
800
700-
600-
500-
400-
:
300-
L = 921(1-2(-0.09(x-0.35)))
Figure 4.
W I I HT I U I I
# , ; ; ; ; ; 10 li z is 14
AGE (YEARS)
i
ź
Estimated von Bertalanffy growth C Ul Iſ Ve and
empirical mean back-calculated lengths at age
with 95% confidence limits.

29
confidence intervals for grids 1608, 1509, 1507, and 1606
We re 0.57 (+0.53), 0.63 (+0.36), 0.66 (+0.48), and
0.45 (+0. 30), respectively. The slopes of the descending
limbs of the catch curves were not significantly different
between grids (f-test, P=0.79). Combining age frequencies
reduced some variability in the descending limb of the catch
curve, however considerable variability remained (Fig. 5).
The total instantaneous mortality rate (Z) and 95%
confidence interval for all grids pooled was 0.67 (+0. 27).
(Regression equations fitted to catch curves are listed in
Appendix D.) Thus, the instantaneous fishing mortality rate
was 0.57, (0.67 less the natural mortality of 0.1). This
rate was partitioned in proportion to the yield to each gear
type to obtain an estimate of the instantaneous fishing
mortality rate for each gear type (Table 6). Using this
information the instantaneous commercial fishing mortality
rate was 0.51 and defined as:
FC as FR + FH + FE.
Yield
Similarities in the growth and mortality rates of
channel catfish from each management area in Saginaw Bay
indicated that fish from the entire Bay comprised a unit
stock for the purposes of a yield-per-recruit analysis
(Gulland 1969). Parameters estimated from pooled growth and
mortality data were used in yield computations. Using the
best estimate of sportfishing mortality (0.06) and natural
mortality (0.1), yield to the commercial fishery per 1000
30
8.8
7- / \ - -
/ |Y = 11.0-0.67(X)
f R2 = 0.86
6- /
--/
Gſ -
/
5-
rº /
O /
º 1
O 4 - ſ
© /
O ?
O /
—l J
3- /
/
/
|
/ \
* /
2 f \
d '.
1-
O U T - I W I I I T T T iſ T
2 3 4. 5 6 7 8 9 10 11 12 13 14
AGE(YEARS)
Figure 5. catch curve with , regression line fitted to the
descending limb (grids combined).

3 1
Table 6. Yield to each gear type, fraction of total yield
attributable to each gear type, and estimates
of the instantaneous fishing mortalities due
to each gear type in 1981.
Fraction of
Yield total In Stantaneous
Gear (kg) yield fishing mortality
Trap nets 146, 626 0. 563 - 0.32
Set hooks 63, 481 0.244 0.14
Seines 20, 809 0. 080 0 - 05
Sport gear 29, 479 0. 1 1 3 0.06
Total 260, 395 s 1. 00 0. 57
recruits was examined as a function of the instantaneous
commercial fishing mortality (FC) and the age at entry to
the fishery (xe), or commercial minimum size limit (le).
Model predictions indicated that yield per 1000 recruits
could be increased by increasing *c and/or decreasing FC
from their existing Or estimated values of 381 mm
(15 inches) and 0.51, respectively (point Q in Fig. 6). The
largest gains in commercial yield could be realized by
increasing *c.
One approach to evaluate the effects of changes in FC
and *c upon commercial yield was to examine each of these
parameters independently as a function of yield per 1000
recruits. At the existing or estimated values of FS (0.06),
M (0.1), and *c (381mm, 15 inches) model predictions
indicated that the maximum commercial yield per 1000
32
22 789.8
20- 400 KG - 763.9
2. d
Cº. 13- 732.9
< 500 KG T
º . - *
j is- -695.8 #
> *
a. 14 - 650 KG T -651.4 5
5 s:
H. 12- ___ H598.2 °
(/) EUMETRIC LINE - " " " T 3
9: >
La- 10- - 534.6 2:
> :
º
Lal cº »
O 8 458. 3
<ſ
6- - 367. 1
4 --——I-------- -—I—, w— I- w—w 257. 9
0.0 0.5 1.0 1.5 2.0
COMMERCIAL INSTANTANEOUS FISHING MORTALITY
Figure 6.
Yield isopleths per 1000 recruits as a function
of FC and x (corresponding values of *c a e
indicated), w? th M= 0 . 1 and FS = 0 . 06. The
commercial fishery was operating at point Q in
1981 .

33
recruits occurred when the value of FC was approximately
0.20 (Fig. 7). Thus, at the estimated FC value of 0.51 the
commercial fishery was growth over fishing (Cushing 1981).
Model predictions indicated that a decrease in FC from 0.51
to 0.20 produced a 5.7% increase in equilibrium yield per
1000 recruits to the commercial fishery and a 1 17.6%
increase in equilibrium yield per 1000 recruits to the sport
fishery. Commercial fishing mortality could be reduced by
reducing the total amount of fishing gear licensed on the
Bay, further restricting areas of the Bay open to commercial
fishing, restricting the length of the fishing season,
imposing taxes, and in other ways (Clark 1976).
To evaluate the implications of gear requests by
commercial fishermen only changes in the total amount of
gear licensed were considered. Thus, an increase or
decrease in FC would be controlled by increasing Or
decreasing the amount of gear licensed. Using this
management framework, a change in FC could be accomplished
in two ways: (1) by proportionally changing the fishing
mortality caused by all commercial gears, or (2) by
selectively changing the fishing mortality caused by any
combination of commercial gear types or any one gear type.
Proportional changes in the instantaneous fishing mortality
caused by each commercial gear type would not change the
distribution of catch among gear types. However, selective
changes in fishing mortalities due to each gear type would
alter the distibution of catch among each gear type. For
34
900
000-
700-
600-
500-
400-
300.
200
100
d
Figure 7.
v- y w- I w wr- w-
v- Ny I w-
0.5 1 1.5 .
COMMERCIAL INSTANTANEOUS FISHING MORTALITY
Yield per 1000 recruits as a function of FC with
FS=0.06, M=0. 1, and l = 38.1 mm (solid line); and
for a range of FS + M values, with c=38 In Iſl
(broken lines).

35
simplicity, the change in yield per 1000 recruits was
examined - for each gear type as a function of the
instantaneous fishing mortality caused by only On e
commercial gear type. A reduction in fishing mortality of a
single commercial gear type reduced the yield to that gear
type and increased yield to other gear types. Conversely,
increases in fishing mortality of a single commercial gear
type increased the yield to that gear and reduced yield to
other gear types (Figs. 8, 9, and 10). Any net reduction in
commercial fishing mortality increased yield t O
sport fishermen.
Considering the commercial fishery as a whole, the
value of FC which produced the maximum yield per 1000
recruits increased as values of FS + M increased (Fig. 7).
Therefore, errors in the estimates of FS, M, or both; or
changes in the values of FS, M, or both could affect the
conclusions derived from this analysis. I examined the
potential consequences of these problems by analyzing the
relation of FC to commercial yield per 1000 recruits for a
range of FS + M values (Table 7). For each FS + M value
examined a total instantaneous mortality rate of 0. 67 was
maintained by adjusting the estimate of FC so that:
F.S + M + FC = 0. 67.
At the FS + M value of 0.25, and corresponding FC
value of 0.42, model predictions indicated that the
commercial fishery was fishing at a rate that produced the
36
200
100 SPORT GEAR
O-º-º-º: 2Zzz 4.
0.0 0.5 1.0 1.5 2.0
200
100 SEINES
ok.<2×z-
0.0 0.5 1.0 1.5 2.0
400
y
C
300 %
200 SET HOOKS
100
O-º-º-º-º-º-º-º-º-4 2Zzz
0.0 0.5 1.0 1.5 2.0
400
w Ty v-
º NETS2%
w o's to 1's
TRAP NET INSTANTANEOUS FISHING MORTALITY
Figure 8. Yield per 1000 recruits to sport gear, seines,
set hooks, and trap nets as a function of FR,
where FS=0.06, FE=0.06, FH=0. 14, M=0. 1, and




37

200-
100 SPORT GEAR
O 2ZZzzzzzzzzzzzzzzzzzz
0.0 0.5 1.0 1.5 2.0
200
d
100- SEINES
* 0–
§ 0.0 0.5 1.0 1.5 2.0
2 400 -
=) d "Z
5 soo- Ž%
# % %
o 200- %5; HOOKS 2
O º
$2 100- % %
ſº % %
ial
On 0 |
O 0.0 0.5 1.0 1.5 2.0
I
> 500
400 º
soo?
% TRAP NETS
200 * -
100 - -
%
0.0 0.5 | 1.0 1.5 2.0 .
SET HOOK INSTANTANEOUS FISHING MORTALITY
Figure 9. Yield per 1000 recruits to sport gear, seines,
set hooks, and trap nets as a function of FH
where FS=0.06, FE=0.05, FR=0.32, M=0. 1, and



e=38 IIMT e
38
200-r-
100 SPORT GEAR
O 2ZZzz 4.
0.0 o's to 1.5 2.0
400
200- SEINES
100- /* %
.k.2%
0.0 0.5 1.0 1.5 2.0
200
º SET HOOKS
0.0 0.5 1.0 1.5 2.0
400




300
> ::
º º TRAP NETS &
2Z.
2.0
O-º-º-º-º-º-º-º-º-º-º-º-º:
0.0 0.5
SEINE INSTANTANEOUS FISHING MORTALITY
Figure 10. Yield per 1000 recruits to sport gear, seines,
set hooks, and trap nets as a function of FE
where FS= 0.06, FH=0. 14, FR=0. 32, M=0. 1, and
le=38 ITIIl e
39
Table 7. Range of FS + M values used to examine the
sensitivity of yield per 1000 recruits as a
function of FC, and range of FS + M (and FC)
values used to examine the sensitivity of yield
per 1000 recruits as a function of lc.
Mortality
parameter (s) * Values
F.S + M 0. 10 0 . 16 0.20 0.25 0.30 0.35 0.40 0.45
FC 0. 57 0. 51 0. 47 0. 42 0 . 37 0.32 0.27 0.22
maximum yield per 1000 recruits. In analyses where FS + M
values were less than 0.25, a growth over fishing condition
existed, which indicated that a decrease in FC would
increase the yield per 1000 recruits. Model predictions
also indicated that where FS + M values were greater than
0.25 a condition existed where commercial fishing mortality
was not great enough obtain the maximum yield, and an
increase in FC would increase the yield per 1000 recruits.
At the estimated FS + M value of 0.16 commercial fishing
mortality was excessive and resulted in a condition of
growth over fishing. However, if FS, M, or both increased or
were underestimated by an amount greater than 0.9 (so that
FS + M was greater than 0.25) model results would favor
granting commercial fishermen an increase in the amount of
gear licensed. Therefore, in situations where an increase
in FC is being considered the importance of accurately
estimating and monitoring changes in FS and M is clearly
demonstrated.
40
At the existing or estimated values of FC (0.51), FS
(0.06), and M (0.1), model predictions indicated that the
maximum commercial yield per 1000 recruits occurred when *c
was approximately 550 mm, 22 inches (Fig. 11). Increasing *c
from 381 mm (15 inches) to 550 mm (22 inches) produced a
28.8% increase in equilibrium yield per 1000 recruits to the
commercial fishery and a 171.6% increase in equilibrium
yield per 1000 recruits to the sport fishery. Increases in
le up to 550 mm (22 inches) resulted in increases in yield
per 1000 recruits to both the sport fishery and commercial
fishery, however when *c became greater than 550 mm
(22 inches), yield to the commercial fishery declined
(Fig. 11).
The value of *c. which produced the maximum yield per
1000 recruits, was also sensitive to changes in the value of
FS + M, and this sensitivity was examined for the range of
values presented in Table 7. At the FS + M value of 0.30,
and corresponding FC value of 0.37, model predictions
indicated that the current minimum commercial size limit of
381 mm (15 inches) produced the maximum yield per 1000
recruits (Fig. 12). In analyses where Fs + M values were
less than 0.3, a condition existed where fish were harvested
at a size less than that which produced the maximum
commercial yield, which indicated that an increase in *c
would increase the yield per 1000 recruits. Model
predictions also indicated that where Fs t M values were
greater than 0.3 a condition existed where fish were
4 1

Figure 11.
700
600 –
sos.
•os.
soo.
ass-
100-
SPORT FISHERY
2.
300
lſ T W T
400 500 600 700
COO

zoo.
•oo-
soo-
•oo-
soo-
zoo.
100-
.
O
300
U I U wr
400 500 600
comMERCIAL MINIMUM SIZE LIMIT (MM)
Yield per 1000 recruits to the
T
700
2.
8 OO
commercial
fishery and sport fishery as a function of *c
where FS= 0.06, FC=0. 51, and M=0. 1.
42
1300
1200-
1100-
1000-
O I U wr- ſ w- =====
300 400 500 600 700 800
COMMERCIAL MINIMUM SIZE LIMIT (MM)
Figure 12. Yield per 1000 recruits as a function of l- with
FC= 0.51, FS=0.06, and M=0. 1 (solid line) ; and
for a range of FS + M values (broken lines),
with FC= 0 , 51.

43
harvested at a size greater than that which produced the
maximum yield, and a decrease in le would increase the yield
- per 1000 recruits. At the estimated FS + M value of 0.16, a
low commercial minimum size limit resulted in a condition of
growth over fishing. However, if either FS, M, or both were
were underestimated or increased by an amount greater than
0.14 (so that FS + M was greater than 0.3) model results
would favor a *c of 381 mm (15 inches), or perhaps less.
Thus, accurately estimating and monitoring changes in FS and
M is also important in situations where a change in *c is
being considered.
DISCUSSION
Age Composition and Size Structure of the Catch
Estimated age distributions of the catch of channel
catfish were different in each grid sampled. Differences in
age distributions could be attributed to age-specific
differences in availability, age-specific differences in
vulnerability to the sampling gear, and real differences in
age composition between grids. Greater proportions of
younger age groups were captured in grids 1608 and 1606
(Table 5). Randolph and clemens (1976) found that smaller
channel catfish occupied shallower water areas in culture
ponds, and larger channel catfish occupied deeper water
areas. Grids 1608 and 1606 are characterized by more
extensive shallow water a reas, perhaps attractive to younger
age groups, and grids 1509 and 1507 are characterized by
more extensive deep water areas, perhaps attractive to older
age groups. This suggests that the observed differences in
age distributions were due to differences in availability of
various age groups. Although it appeared that younger age
groups were captured in proportion to their availability in
grids 1608 and 1606, even larger proportions were retained
in nets in grid 1606, which had smaller pot mesh than nets
in grids 1608, 1509, and 1507. This suggests that some age-
specific differences in vulnerability to the fishing gear
also existed.
44
45
Growth in Length
There were significant differences in the growth rate
of channel catfish between sexes in grids 1509 and 1608.
Mean back-calculated lengths at age were greater for males
in grid 1509 and greater for females in grid 1608
(Appendix A). Few studies have compared growth in length of
channel catfish between sexes. Elrod (1974) and Ambrose and
Brown (1971) contended that growth in length between sexes
was similar enough in their studies to be combined. DeRoth
(1965) found that the average back-calculated lengths of
males and females in Lake Erie was similar up to age 4.
After age 4, mean back-calculated lengths at age were greater
for males. In experimental ponds, Beaver et al. (1966)
found that the average length of males was greater than the
average length of females. Thus, growth differences between
sexes in this study, in particular in grid 1608, were more
likely due to differential vulnerability of males and
females to capture rather than real growth differences.
Grids 1509 and 1608 were sampled near the spawning season at
a time when behavioral differences have been observed
between sexes. Trap nets are passive sampling devices, and
vulnerability to capture was dependent upon In OV ement .
Larger channel catfish spawn earlier in the spawning season,
and males drive away females and care for the eggs after
spawning (Clemens and Sneed 1957). Thus, larger males may
have been less vulnerable to capture in grid 1608, if
sampling took place early in the spawning season when larger
46
males were on their nests. Also, hatching has been observed
to take place after approximately 1 week of incubation
(Clemens and Sneed 1957). If adult males disperse soon
after the eggs hatch, it is plausible that larger males were
more vulnerable to capture at this time. This may explain
the larger back-calculated lengths at age observed for males
in grid 1509, although growth differences evident in grid
1509 were consistent with other reported results.
When sexes were pooled, no significant differences in
growth rate were detected between grids. However, mean
back-calculated lengths at age were less for channel catfish
from grids 1608 and 1606 than from other grids. These
differences may be real, may be associated with age-specific
differences in availability, or may be associated with net
selectivity. More younger fish were captured in grids 1608
and 1606 than in grids 1509 and 1507 (Table 5). In grid
1608, a greater proportion of smaller channel catfish may
have been captured, because larger fish, in particular
males, were probably on or near nests and less available at
the time of sampling. In grid 1606, the smaller pot mesh
utilized may have selected greater proportions of smaller
fish. Although mean back-calculated lengths at age were
smaller in grids 1608 and 1606 than in other grids, In O
significant differences in growth rates or intercepts were
detected between grids when regression equations were f it to
lengths after age 6, the age of complete recruitment to the
commercial fishery.
47
Pooled mean back-calculated lengths at age in this
study, were about average when compared to those reported
from other northern regions, and were smaller than those
reported from more southerly locations (Table 8). Mean
back-calculated lengths in this study were less than those
reported 10 years earlier by Eshenroder and Haas (1974). In
their study, channel catfish were captured with commercial
and experimental nets; experimental nets captured younger
age classes of channel catfish in greater proportion than
commercial gear types (Eshen roder and Haas 1974). I sampled
fish from commercial trap nets exclusively. If differences
in selectivity of the gear caused the observed growth
differences, one would expect fish in the present study to
be larger on the average at each age, since commercial nets
probably selected greater proportions of larger fish.
However, mean back-calculated lengths at age were greater in
the study conducted by Eshenroder and Haas (1974), thus,
gear selectivity probably did not account for the observed
growth differences.
Physical, chemical, and biological conditions have
been demonstrated to be quite dynamic in Saginaw Bay
(Rossman and Treese 1982), and these changes may explain the
observed differences in growth. Conditions which may have
caused a decrease in growth of channel catfish include:
§
Table 8. Mean back-calculated lengths (mm) at age of channel cat f ish from Saginaw Bay in 1981
(with 95% conf idence intervals) and 1971 .
and from other selected locations.
Age
Year (s) of Sample
locat i on study size 1 2 3 4 5 6 7 8 9 1O 1 1 12 13 1.4
Sag i naw Bay. Lake Huron 198 || 9 || 6 54 132 198 256 31 O 358 42O 469 507 546 594 6O4 612
+ 1 + 2 + 2 + 2 + 3 + 3 +5 + 6 + 6 + 8 + 1 4 + 2C) + 76
Sag i naw Bay, Lake Huron " " ' 1971 253 69 157 254 335 4O4 490 56 569 589 6O7 63O
Lake Erie, Western Bas in ' " ' 1965 1,478 63 165 226 269 297 33o 363
Lake Erie, Michigan Waters " " ' 1971 495 17O 22 1 264 3O5 340 373 386 4 17 452 505 549
Lake St. Cl a i r ' ' ' 1969–71 5O7 76 208 226 272 350 383 434 485 53 1 564 6O4
St. Lawrence River ' " ' 1975 28 1 is 164 204 239 272 302 33o 353 377 4oo 432 455
Lake Sharpe, South Dakota * ' 1945-56 535 46 124 196 256 3. 1 2 38 1 442 49 O 546 6 17 645 640 676
Oklahoma ' " ' 1946-54 7.717 1O || 2 15 3O2 368 409 452 504 555 6.O7 63O 644 648 655 73 1
Santee-Cooper Reserveſ on System,
South Carol na * * 1959 21 O 86 85 28 4 368 4 42 53 1 SO2 665 726 772 807 853 9 7 9C)4
Farm Ponds, Central Texas ' ' ' 1972 82 1 78 333 429 5 16
' ' ' Eshenroder and Haas (1974)
' ' ' DeRoth (1965)
' ' ' Magu in and Fradette (1975)
' ' ' El rod (1974)
" * ' Jenkins (1954 )
' ' ' Stevens (1959)
' ' ' Prent ice and Whites i de (1975)
49
(1) A decrease in primary productivity as demonstrated
by a decrease in median chlorophyll a levels from
1974 to 1979 (International Joint Commission 1980).
(2) Intraspecific competition due to an increase in
channel catfish abundance as demonstated by the
increases in catch per unit of effort (Fig. 2).
(3) Interspecific competition due to intensified
salmonid and walleye Stizostedion vitreum stocking
programs.
Numerous other factors could contribute to the observed
decrease in mean length at age.
Weight-Length Relation
Differences in the linearized weight-length relations
between management areas probably reflect seasonal change in
weight and length (Bagenal 1978). Grids were not sampled
during the same time period and the slope of the relations
decreased from spring to summer (Appendix C). This
indicated that condition or plumpness decreased (Lagler
1956). Simco and Cross (1966) noted a similar decline in
condition of channel catfish in the late summer. Despite
these seasonal changes in condition, pooled weight and
length data were used in yield analyses to calculate the
approximate weight of the catch.
Total Mortality Rate
Ricker (1975) identified the following conditions or
assumptions implied in interpreting the descending right
limb of a catch curve:
50
(1) The age groups under study are equal in number when
each is recruited to the fishery.
(2) Survival rate is uniform with age.
(3) Fishing and natural mortality rates are uniform with
age, since both comprise the total mortality rate
which is the complement of the survival rate.
(4) The sample is random.
Catch curves for each grid were characterized by
irregular right limbs which were probably caused by variable
recruitment. The right limbs exhibited the usual decreasing
trend with age and the variability appeared to be random.
Thus, I assumed that the slope of a line fitted by least
squares to the descending limb adequately estimated the
total instantaneous mortality rate. The irregularity of the
right limbs of the catch curves also precluded detection of
curvature, useful in evaluating the assumption of uniform
survival. However, commercial fishing effort information
provided insight to the constancy of survival, since
commercial fishing accounts for the bulk of total mortality.
Commercial fishing effort, and thus commercial fishing
mortality were relatively constant in years just prior to
1981, the refore survival rate was also probably C On Stant
(Fig. 2, B).
The assumption of random sampling implies that
differences did not exist in age-specific vulnerability to
trap net S after the age of maximum vulnerability.
Differences in age structure of the catch were detected
between grids, however no differences in total mortality
51
were detected after the age of complete recruitment. This
indicated that differences in age composition were primarily
due to differences in availability or selectivity before the
age of complete recruitment to the fishing gear. Latta
(1959), and Laarman and Ryckman (1982) demonstrated that
trap nets were size selective for a number of species.
Ricker (1975) contended that error from catch curve
mortality estimates due to size-specific vulnerability was
unlikely to be large for trap nets, but encouraged efforts
to assess selectivity. Yeh (1977) showed that trap nets
were more effective in representing the length class
frequency of a population of channel catfish than gill nets
and hoop nets in large inland lakes in Texas. Assessment of
the selectivity of commercial trap nets for channel catfish
and collection of age frequency data over a series of years,
to minimize the effects of variable recruitment, would
enhance the reliability of the total mortality estimate.
The total instantaneous mortality rate for channel
catfish estimated in this study was identical to the value
estimated by Eshenroder and Haas (1974). This value also
lies within the range of other published values (Table 9).
Thus, it appears that the physical, chemical, and biological
changes that have occurred since the study by Eshenroder and
Haas (1974), have had little effect on total mortality.
Fishing Mortality. Rates
The method used to estimate the instantaneous fishing
mortality rate for each gear or fishery was contingent upon
52
Table 9. Instantaneous total mortality rates estimated for
Saginaw Bay in 1981 (with 95% confidence
interval) and 1971, and for other selected
locations.
Year (s)
- of
Location study Z
Saginaw Bay, Lake Huron 1981 0.67 (+0. 27)
Saginaw Bay, Lake Huron ‘’’ 1971 0.67
Des Moines River, Iowa “*” 1966-69 0 - 637
Upper Mississippi, Iowa ‘’’ 1977-79 0.94
Lake Sharpe, South Dakota' * * 1945-56 0.37
Rivers of Sacramento Valley, 1955–59 0.82
California ( * *
' ' ' Es&hen roder and Haas (1974)
( * 1 Mayhew (1972)
( * Pitlo and Bonneau (1979)
( * } Elrod (1974)
( * McCammon and La Faunce (1961)
the reliability of the catch statistics used to partition
FT. Commercial records provided the catch for each gear
type as reported by commercial fishermen. These records
were assumed to be reliable. Sport fishing post card survey
estimates, however, have been demonstrated to be inflated,
and corrections for this have been described. These
correction factors were largely derived from exploitation
studies of lake trout Salvelinus namaycush and yellow perch.
Specific corrections for channel catfish would improve the
53
reliability of sport harvest estimates used here, and thus
improve the estimates of the instantaneous fishing rates for
each gear or fishery. A sport fishing mortality estimate of
approximately 0.20 would result if the estimate of
sport fishing yield were not corrected. Coupling this FS
value with the best estimate of M (0.1) yielded an FS + M
value of 0.30. Using this FS + M value model predictions
indicated that the commercial fishery was operating at a
minimum size limit and fishing rate very near that which
produced the maximum yield per 1000 recruits (Fig. 12).
Thus, results of the yield analysis greatly depend upon the
correction factor used in the estimation of sport fishermen
yield. An inaccurate correction factor could change the
conclusions drawn from this analysis.
Ultimately, the reliability of all of these fishing
mortality estimates depend upon the accuracy of the estimate
of F.T. The indirect procedure used to estimate FT in
Saginaw Bay was exclusively dependent upon the estimate of Z
and the value of M estimated by Eshen roder and Haas (1974).
The degree to which FT and M make up Z can significantly
influence yield computations. Model results indicated that
when natural mortality accounted for a larger percentage of
total mortality; maximum yield was obtained at greater
fishing rates (Fig. 7) and/or lower minimum size limits
(Fig. 12). Estimates of M and FT from the Sacramento Valley
were 0.38 and 0.44, respectively (McCammon and LaFaunce
1961); each mortality parameter constituted approximately
54
one half of Z. In contrast, values from Saginaw Bay were
0.1 and 0.57 for M and FT, respectively. Here, the
instantaneous fishing mortality constituted a significantly
larger portion of Z, and model predictions indicated that
maximum yield was obtained at a relatively low fishing rate
and high minimum size limit. A more direct estimation
procedure to assess fishing and natural mortality (such as a
tag-recapture procedure) would ultimately be necessary to
improve estimates of natural and fishing mortalities and/or
validate the method used to estimate fishing mortalities.
Yield
Beverton and Holt (1957) listed the following
assumptions implied in a cohort yield-per-recruit model:
(1) Yields predicted are those which exist under
equilibrium conditions.
(2) Extrapolation of model results to a population
requires constant recruitment.
(3) Rates of growth and natural mortality remain
constant in response to other changes.
(4) All fish older than age x
are equally vulnerable
to capture. *ms
r
Ricker (1975) identified the following assumptions
applicable to the competing gear types modeled in the
present study:
(5) The units of gear are distributed such that all
fish are equally vulnerable to capture and there is
no possibility of localized depletion of the stock.
(6) There is no physical interference among gear types
with respect to their operation.
The utility of the results of this analysis depend to a
55
large degree, upon how well all of these assumptions were
met . with respect to the first assumption it is important
to recognize that predictions from this analysis in response
to a particular management action will not be realized
immediately. Lag time to attainment of a new equilibrium
will depend upon the management action, the generation time,
and recruitment (Walters 1969). For example, an increase in
the commercial minimum size limit would ultimately increase
yield, yet upon implementation a short-term reduction in
yield to the commercial fishery would result. Clark (1976)
discusses strategies which minimize these short term losses.
The impacts of management actions upon recruitment
cannot be assessed with this yield-per-recruit model.
Although recruitment has exhibited some variability, the
history of stable landings indicates that the assumption of
constant recruitment is probably appropriate for the current
total mortality rate and growth regime. However, fishing
mortality could become great enough to affect recruitment.
Russell (1942) documents some devastating consequences of
recruitment over fishing in the North Sea, and symptoms of
recruitment over fishing have been documented for channel
catfish in the Mississippi River (Pitlow and Bonneau 1979).
An understanding of the relation between stock and
recruitment would significantly enhance management
capability.
Changes in growth and natural mortality in response
to changes in management actions or environmental changes
56
have not been evaluated, yet model predictions hinge upon
the constancy of these vital statistics. Using the best
mortality and growth parameter estimates, model predictions
indicated that an increase in the minimum commercial size
limit and/or reduction in commercial fishing mortality would
increase both commercial and sportfishermen yield. Either
of these changes would ultimately change the relative
abundance of some age classes of channel catfish (Ricker
1975). These density changes could affect growth and
natural mortality. Model results indicated that small
changes in natural mortality significantly affected model
predictions (Figs. 7 and 12). However, the apparent
increase in abundance of channel catfish and changes in
environmental conditions, since the study by Eshenroder and
Haas (1974), have not resulted in any apparent changes in
natural mortality as evidenced by the constancy of total
mortality and fishing effort from 1971 to 1981. Growth,
however, may change in response to changing environmental
conditions or changes in density as a consequence of a
management action. The growth of channel catfish in Saginaw
Bay in 1971 and in Michigan waters of Lake Erie were used to
study model predictions under different growth regimes.
Mean back-calculated lengths at age of channel catfish in
Saginaw Bay in 1971 were greater than those in the present
study and mean lengths at age after age 6 were less in
Michigan waters of Lake Erie than in the present study
(Table 8). Using the 1971 growth regime in a yield analysis
57
similar to that performed in the present study, Eshenrod ser
and Haas (1974) concluded that an increase in the minimum
commercial size limit would increase yield. I concluded the
same using von Bertalanffy parameters fit to Lake Erie
growth data with the Saginaw Bay mortality regime. Thus, it
appears that small changes in growth that may occur in
response to a management action, or change in environmental
conditions, will not significantly affect model predictions.
The assumption of equal vulnerability to each gear
type was not directly assessed. Not all areas of Saginaw
Bay are commercially fishable, yet all areas are accessable
by sport fishermen. Each commercial gear type is suited to
exploit fish at a particular depth, on a particular bottom
type, and to some degree during a particular season. Thus,
all fish were probably not equally vulnerable to all gear
types. Since it appears that fishing does not result in
localized depletions of the stock, model bias due to
differences in vulnerability would probably be small .
However, a better understanding of the characteristics of
the catch from each gear would permit refinement of the
model and enhance predictions. The nature of the
distribution of the different types of fishing gear in time
and space indicates that physical competition between gear
types would be minimal.
Although the assumptions of the yield-per-recruit
model cannot be rigidly met, some useful management insights
can be gleaned from this study. First, examination of gear
58
competition in a realistic management framework Wa S
interesting. With the best available growth and mortality
estimates, I determined that increases in commercial fishing
mortality through the licensing of additional gear, would
not increase yield to the fishery as a whole. Each
commercial gear type was operating at a level which produced
less than the maximum yield per 1000 recruits to that
particular gear (Figs. 8, 9, and 10), thus, yield to any one
gear type could be increased by increasing fishing mortality
due to that particular gear type. These increases, however
could only be realized with accompanying losses to other
gear types and a net loss to the fishery as a whole.
Second, yield-per-recruit analyses of the commercial fishery
indicated a condition of growth over fishing. This condition
could be corrected by increasing the minimum commercial size
limit, reducing the commercial fishing mortality, or a
combination of both. The largest gains in yield could be
obtained by increasing the minimum commercial size limit
(Figs. 6).
Increasing the minimum commercial size limit would
appear to benefit both the commercial and sport fishery.
Although all model results presented suggest that catfish
yield could be increase by increasing the commercial minimum
size limit, the tenuous nature of the mortality parameter
estimates precludes any immediate recommendation for a
change. However, efforts should be made to more accurately
estimate mortality parameters, and entry should remain
59
limited until these assessments are completed and specific
management plans can be formulated.
LITERATURE CITED
Allen, K. R. 1966. Determination of age distribution from
age-length keys and length distributions, IBM 7090,
7040 FORTRAN V. Transactions of the American Fisheries
Society 95: 230-231.
Ambrose, J., Jr., and, B. E. Brown. 1971. Fecundity, age,
growth, and condition of channel catfish in an Oklahoma
reservior. Proceedings of the Oklahoma Academy of
Science 5.1 : 15-22.
Bagenal, T. 1978. Methods for a SS eSSIſlent of fish
production in fresh water, IPB Handbook No. 3.
Blackwell Scientific Publications, London, England.
Baldwin, N. S. , R. W. Saa feld, M. A. Ross, and H. J.
Buettner . 1979. Commercial fish production in the
Great Lakes 1867-1977. Great Lakes Fishery Commission,
Technical Report No. 3, Ann Arbor, Michigan, USA.
Beaver, J. A., K. E. Sneed, and H. K. Dupree. 1966. The
difference in growth of male and female channel catfish
in hatchery ponds. The Progressive Fish Culturist
28 : 47-50.
Beeton, A. M., S. H. Smith, and F. F. Hooper. 1967.
Physical limnology of Saginaw Bay, Lake Huron. Great
Lakes Fishery Commission Technical Report No. 12, Ann
Arbor, Michigan, USA.
Beverton, R. J. H., and, S. J. Holt. 1957. On the dynamics
of exploited fish populations. Fishery Investigation
Series 2, Volume 19, Ministry of Agriculture,
Fisheries, and Food (Great Britain).
Bishop, R. C. , and K. C. Samples. 1980. Sport and
commercial fishing conflicts: a theoretical analysis.
Journal of Environmental Economics and Management
7 : 220-233.
Blumer, L. S. 1982. Parental care and reproductive ecology
of the North American catfish Ictalurus nebulosus.
Doctoral dissertation. The University of Michigan, Ann
Arbor, Michigan, USA. -
Chugunova, N. I. 1959. Handbook for the study of age and
growth of fishes. Akademiya Nauk Press, Moscow, USSR.
(English version 1963. Age and growth studies in
fish. Office of Technical Services Washington, D. C.,
USA)
Clark, C. W. 1976. Mathematical bioeconomics. Wiley
Inter science, New York, New York, USA.
60
61
Clark, R. D., Jr., and B. Huang. 1982. The dynamics of
competition between sport and commercial fishing:
effects on rehabilitation of lake trout in Lake
Michigan. Michigan Department of Natural Resources,
Fisheries Research Report 1909, Ann Arbor, Michigan,
USA.
Clemmens, H. P. , and K. E. Sneed. 1957. The spawning
behavior of the channel catfish Ictalurus punctatus.
United States Department of the Interior Fish and
Wildlife Service, Special Scientific Report No. 219.
Washington, D.C. USA.
Cushing D. H. 1981. Fishery biology a study in population
dynamics. The University of Wisconsin Press, Madison,
Wisconsin, USA.
DeRoth, G. C. 1965. Age and growth studies of channel
catfish in western Lake Erie. Journal of Wildlife
Management 29: 280-286.
Elrod J. H. 1974. Abundance, growth, survival, and
maturation of channel catfish in Lake Sharpe, South
Dakota. Transactions of the American Fisheries Society
1 03: 53-58.
Eshen roder, R. L., and, R. C. Haas. 1974. Status of
selected stocks in Lake Huron and recommendations for
commercial harvest. Michigan Department of Natural
Resources, Technical Report 73-32. Pages 105-206 in
Bails, J. D., and M. H. Patriarche, editors. Status of
Selected fish stocks in Michigan's Great Lakes waters
and recommendations for commercial harvest. Michigan
Department of Natural Resources, Fisheries Division,
Lansing, Michigan, USA.
Everhart, W. H. , and W. D. Youngs. 1981. Principles of
fishery science. Cornell University Press, I thaca, New
York, USA.
Fox, D. J., and K. E. Guire. 1976. Documentation for MIDAS
3rd edition. Statistical Research Laboratory, Ann
Arbor, Michigan, USA.
Francis, G. R. , J. J. Magnuson, and H. A. Regier. 1979.
Rehabilitating Great Lakes ecosystems. Great Lakes
Fishery Commission, Technical Report No. 37, Ann Arbor,
Michigan, USA.
Gallucci, V. F. , and T. J. Quinn II . . 1979.
Reparameterizing, fitting , and testing a simple growth
model. Transactions of the American Fisheries Society
108 : 14-25.
62
Great Lakes Basin Commission. 1975. Great Lakes Basin
Framework Study Appendix 8 Fish. Great Lakes Basin
Commission, Ann Arbor, Michigan, USA.
Great Lakes Fishery Commission. 1971 a . Minutes of Lake
Huron Committee Meetings Appendix II. Great Lakes
Fishery Commission, Ann Arbor, Michigan, USA.
Great Lakes Fishery Commission. 1971b. Minutes of Lake Erie
Committee Meetings. Great Lakes Fishery Commission,
Ann Arbor, Michigan, USA.
Gulland, J. A. 1969. Manual of methods for fish stock
assessment. Part 1. Fish population analysis. Food
and Agriculture Organization of the United Nations
Manuals in Fisheries Science 4, Rome, Italy.
Hile, R. 1959. Fluctuations in the commercial fisheries of
Saginaw Bay 1885-1956. United States Department of
Interior, United States Fish and Wildlife Service
Research Report 51, Washington, D. C., USA.
Hile, R. 1962. Collection and analysis of commercial
fishery statistics in the Great Lakes. Great Lakes
Fishery Commmission, Technical Report No. 5, Ann Arbor,
Michigan,
International Joint Commission. 1980. 1980 report on Great
Lakes water quality. Toronto, Ontario, Canada.
Jenkins, R. M. 1954. Growth of channel catfish in Oklahoma
waters: 1954 revision. Oklahoma Fisheries Research
Laboratory Report 41, Norman, Oklahoma, USA.
Kimura, D. K. 1977. Statistical assessment of the age-
length key. Journal of the Fisheries Research Board of
Canada 34:317-324.
Kingsley, C. S., V. F. Gallucci, and T. J. Quinn II. 1980.
Von Bertalanffy growth parameters. Transactions of the
American Fisheries Society 109:252-253.
Laarman, P. W., and J. R. Ryckman. 1982. Relative size
selectivity of trap nets for eight species of fish.
North American Journal of Fisheries Management. 2: 33-37.
Lagler, K. F. 1956. Freshwater Fishery Biology. Wm. C.
Brown Company, Dubique, Iowa, USA.
Latta, W. C. 1959. Significance of trap net selectivity in
estimating fish population statistics. Papers of the
Michigan Academy of Science, Arts, and Letters
44 : 123-138.
63
Low, R. A. , Jr. 1982. Mortality rates and management
strategies for black sea bass off the southeast coast.
North American Journal of Fisheries Management
1 : 93- 103.
Maguin, E., and G. Fradette. 1975. Growth of the channel
catfish (Ictalurus punctatus) in the St. Lawrence River
near Quebec. Journal of the Fisheries Research Board
of Canada 32: 1867-1870.
Marzolf, R. C. 1955. Use of pectoral spines and vertebrae
for determining age and rate of growth of the channel
catfish. Journal of Wildlife Management 19: 243-249.
Mayhew, J. 1972. Some biological characteristics of a
channel catfish population in the lower Des Moines
River with an evaluation of potential commercial
harvest. Iowas Conservation Commission Fisheries
Section, Technical Series 72-2, Des Moines, Iowa, USA.
McCammon, G. W. , and D. A. LaFaunce. 1961. Mortality rates
and movement in the channel catfish population of the
Sacramento Valley. California Fish and Game 47:5-26.
McHugh, J. L. 1980. Progress and problems in implementing
the Fishery Conservation and Management Act of 1976.
Transactions of the North American Wildlife and Natural
Resources Conference 45: 149-156.
Muncy, R. J. 1959. Age and growth of channel catfish from
the Des Moines River, Boone County, Iowa, 1955 and
1956. Iowa State Journal of Science 34: 127-137.
Pitlow J., Jr., and D. Bonneau. 1979. Assessment of change
in commercial size limit of channel catfish. I Owa
Conservation Commission Fisheries Section, Final Report
Commercial Fisheries Investigations, Des Moines Iowa,
USA.
Prent ice, J. A. , and B. G. Whiteside. 1974. Validation of
aging techniques for largemouth bass and channel
catfish in central Texas farm ponds. Proceedings of
the Annual Conference Southeastern Association of Game
and Fish Commissioners 28: 4 14-428.
Rafail, S. Z. 1973. A simple and precise method for
fitting a von Bertalanffy growth curve. Marine Biology
19 : 354-358. *
Randolph, K. W., and H. P. Clemens. 1976. Home areas and
swimways in channel catfish culture ponds.
Transactions of the American Fisheries Society
1 05: 725-7 30.
64
Ricker, W. E. 1975. Computation and interpretation of
biological statistics of fish populations. Fisheries
Research Board of Canada Bullet in 191, Ottawa, Canada.
Rossmann, R., and T. Treese. 1982. Lake Huron bibliography
with limited summaries. Great Lakes Research Division,
Great Lakes and Marine Waters Center, The University of
Michigan Special Report 88, Ann Arbor, Michigan, USA.
Russell, E. S. 1942. The over fishing problem. Cambridge
University Press, London, England.
Rybicki, R. W., and M. Keller. 1978. The lake trout
resource in Michigan Water S of Lake Michigan,
1970- 1976. Michigan Department of Natural Resources,
Fisheries Division, Fisheries Research Report
No. 1863, Ann Arbor, Michigan, USA.
Simco, B. A., and F. B. Cross. 1966. Factors affecting
growth and production of channel catfish Ictalurus
punctatus. University of Kansas, Museum of Natural
History 17:191-256. -
Sneed, K. E. 1951. A method for calculating growth of
channel catfish, Ictalurus lacustris puncta tu S.
Transactions of the American Fisheries Society
80 : 174 - 183.
Sokal, R. R. , and F. J. Rohlf. 1981. Biometry: the
principles and practice of statistics in biological
research 2nd edition. W. H. Freeman, San Francisco,
California, USA. -
Statistical Research Laboratory. 1976. CHI TAB general
description; The University of Michigan, Ann Arbor,
Mishigan, mimeo, 2p.
Stevens, R. E. 1959. The white and channel catfishes of
the Santee-Cooper Reservior and tail race sanctuary.
Proceedings of the Annual Conference Southeastern
Association of Game and Fish Commissioners 13:203-219.
Talhelm, D. R. , R. C. Bishop, K. W. Cox, W. Smith, D. N.
Steinnes, and A. L. W. Tuomi. 1979. Current estimates
of Great Lakes fisheries values: 1979 status report.
Great Lakes Fishery Commission, Ann Arbor, Michigan,
mimeo report 79- 1: 17 p.
Talhelm, D. R. 1979. Fisheries: dollars and cents. Water
Spectrum 11: 8-16.
Walters, C. J. 1969. A generalized computer simulation
model for fish population studies. Transactions of the
American Fisheries Society 98: 505-5 12.
65
Yeh, C. F. 1977. Relatively selectivity of fishing gear
used in a large reservoir in Texas. Transactions of
the American Fisheries Society 106:309-313.
APPENDIX A.
Mean back-calculated lengths at age for both sexes in each
grid and for sexes combined.
66
67
( ! )( v )( L \ )( £ € )( ! Þ )( Z9 )( ! ! ! )( g ） ! )( € L \ )( ZOZ)( ţ) CZ)( 9€Œ )
• • •º /. £6 || 3:9 !! Ț9 į 3F? !! ！!！8 ±93;? !€ ±Z Ț |Z Ț
GO9899€ Z9! O9O AL țy! € ț7998LOC€ 90ț76 }! € }99pºu ! CſuuO O
( ! )( C )( z \ )( 9 ! )(OZ)( 6Č )( G9)( į 9 )( ţ) 8 )(OO ! )( G \ ! )( 9 į į )
• • •! 63FZ Z +!723!} & \;9 !! Ț! !! %±83;93R? ！!！€ £€ £
9O99 199 !! 99Ogț» AL £7! ¿ ?O9€.6OE€ 90º 6 įį € }| 99el euæ 49O 9 ||
• • •• • •( 9 )( g \ )( 6 ) )( ! ? )( ? 9 )( 29)( 98 )(66),( 9 !! !! )( ! ! ! )
O93;9：23;țy Z {į Č Ę! !! ！!63:93;º £€ £€ £
• •1€996țy19ț»Z Z țyZ90ÞO€€9 Z? 6 }! € !ſy G3 | ew
(Z )( 6 )( g ） )( 8 º )( ſºg)( 69 )( Þ G \ )( 99 || )( 16 ! )( gOZ )( 9OZ )( 9OZ )
! € £ 3;8€3;:Z Z {? ! ！;C || 3:$ $ 3;A1 +93;? ?? ！!!€ £€ £
969669O99! !! 9º 1?! € ſºO99! !! C89 ºCOZ! € !69peu ! QuO3)
( ! )( €)(O ! )( ! Z)( ºz. )(ZC)(OL)( 91 )( 18 )( 3 6 )( į 6 )( į 6 )
• • ►89 || 3;! ſº ſ€¢ £OZ. 3;6 | 3:į į 3;63;L+93;93;ț73;
| Z99：299999 !! 99 AL 17OZ ſºO98）! !! €89 CZOZ6€ į69e 1 2u3 -JLOS $
( $ )( 9 )(Z ! )(CZ)( 9?)( ZC)( 81 )( w 8 )( ZO ! )( 9O4 )( 9O4 )( 9O ! )
• • •993;€Z 3;6 | 356 } ）;9! 3:63;A1 +§ 3;93;? ！!!£ 3;
O 19989Ź99€ į 961. №9： ?! 9€.$ $ €890! OZ9€ !693 į ſew
( C )( © 1 )( £ € )(OL)(O6)(G6)( € \ ! )(O9 ! )( ! 9 !! )(Oļ Z )( O || Z. )(Oļ Z )( O ). Z )
Z Ź Ź ŻĘAL Z 3F9 !! Țț» į £Z ! !;! !! ！!！63;:A1+93;º £tº £? ?Z3;
Z ! 9LO9�O99999 į 991ț7AL Z Þ�991 | €29?ZOZ9€ !89peu ! QuJO O
( z )( 8 )( 8 ! )( ! ? )( €9 )( G9)( G9)( 86 )( 66 )( 8O ! )( 9O4 )( 9O4 )( 9O4 )
Z Z Z +1& +6 | 3:6!} +9 į 3;9} +& \ 3;63:93;93;93;9 ±C ::
© į 9Z69€8910916?! 9ț7! ¿ ?69€.C ! E69 Z! OZ9€ !19º 1 eurº 46O9 !
→ ∞， ∞( v )( g \ )( 82 )(9€)( 6€)( L º )( 81 )( 81 )(96)(96)(96)(96)
9 13:ſ€Œ Œ6 !! ȚG! 3*9 !! ！!！9 !! ！!O ! ！;63:93;93;93;tº 3;
1629O€9! 89! ? £96 ſº�8. ſyO 10Z Zº992?OZ1€ !693 i ew
( g )( 8 ! )( 19 )(O6)( 86 )(9 ! ! )(6Oz)(Z \ Z)( į 92 )( 99Z )( 99?)( G9-Z )
€ AL3F9€3:? !! ！!! !! 3:63:A13，$3;93;ţ» +€ £€ £Z Zſ
169€ 89º £79OO969 ſºC || ?Z909OCC9?€6 ||9.^ !9 ſºpeu ! QuJO O
( v )( 8 )( 8?)( 10 )(z tº)(Og)( 86 )(OO! )( $ && ! )(ZZ ! )( € Z ! )( CZ || )
983;© ț¢ £8 || 3;9 į 3;9 !! ！!！© || 3:03;Al3F93;93;§ 3;€ £
€ £ 9§ 696 º 9& ! 99999 $ ſº8ºCZO9! 90£6||9€ ſ9 ſºĐ į eurº -I9O9 į
( ! )(O! )( ÞC)(9ț»)( 69 )( 89 )( ÞO! )( GO! )( CC 1 )(9€ ! )( 98 % )(9€ į )
• • ►! 9†© ® È9! 3;? !! ！!O ! ！Al 3;AL3F§3;? ?? ?€ 3;
�£ €9Č 19£ €968ț»Zgº6Oſº9988Oºº gº£6}9.^ !AL £7Đ || ew
€ !Z !! !O !68L99�€Z!|D | J5)
æ6y
( * ez , se į duespeu ! quuOo Đų ą Oļ uns ! OuKewuse į eue 4 pue se į eu 3 O sez ! Sa į duesĐuļļ ĐJO 43 Jºu|}
º sueu ! oedsl l e Jo 4 peu ! uue Q ep you seraxaS° u 36uæ !uļoeæ JO Į pēļ e3 ſ pu į 3 4 e 1 e^-13 } u \, e ou3p || Juo3%96pue
ez , s e i dues )• peu ! quoo sexæs Jog pue p | J6 uoea u į sexas uņoq Jog æ6e 3 e (uu) sųą6uæ ! p3) e i no 1 e3–>19eq ueæW ' ! – w e I qe 1
68
APPENDIX B .
Table B-1. Polynomials fitted to back-calculated lengths at
all ages for each sex in each grid and for
sexes pooled within grids. (Sample sizes and
coefficients of determination are indicated.
Sex was not determined for for all
specimens, therefore the sample sizes of
males and females may not sum to the
combined sample size.)
Regression 2
Grid equation N R
Male - -23.7 77.8 (x) - 2.3 (x*) 945 0.95
1608 Female - - 18.7 72.9 (x) - 1.6(x*) 856 0.95
combined y - -20.8 + 75.3 (x) - 1.9(x*) 1869 0.95
Male - -8.8 + 73.9 (x) - 1.5(x”) 709 0.94
1509 Female - -10.4 + 75.0(x) - 2.0 (x*) 871 0.93
combined y = -10.7 - 75.1(x) - 1.9(x*) 1616 0.94
Male - -8.5 + 73.3 (x) - 1.7 (x*) 681 0.95
1507 Female - -3.1 + 70.8 (x) - 1.5(x”) 597 0.94
combined y - -6.1 + 72.5(x) - 1.7 (x*) 1339 0.95
Male - - 14.5 + 74.3 (x) - 1.9(x*) 592 0.95
1606 Female - - 16.9 + 76.3 (x) - 2. 1 (x*) 612 0.96
combined Y - - 16.4 - 75.8 (x) - 2. 1 (x*) 1229 0.96
69
APPENDIX C.
Table C-1. Linearized weight-length regressions for each
sex in each grid, for sexes pooled within
grids, and for grids and sexes pooled. (Sample
sizes and coefficients of determination are
indicated. Sex was not determined for all
specimens, therefore sample sizes of males
and females may not sum to the combined
sample size.)
Regression 2
Grid equation N R
Male Y = -21. 6 + 3.5 (X) 135 0.98
1608 Female Y = -2 1 - 3 + 3. 5 (X) 125 0 - 99
Combined Y = -2 1.5 + 3.5 (X) 267 0.98
Male Y = -20. 5 + 3. 3 (X) 96 0.99
1509 Female Y = -2 1 - 0 + 3.4 (X) 109 0.98
Combined Y = -20. 8 + 3.4 (X) 21 1 0 - 99
Male Y = -20. 7 + 3. 3 (X) 106 0.97
1507 Female Y = -20. 8 + 3.4 (X) 92 0.96
Combined Y = -20. 8 + 3. 4 (X) 206 0.97
Male Y = - 19.8 + 3. 2 (X) 117 0.99
1606 Female Y = - 19 . 8 + 3. 2 (X) 1 16 0.99
Combined Y = - 19.8 + 3. 2 (X) 236 0.99
70
APPENDIX D.
Table D-1. Equations fitted to the descending limbs of
catch curves for each grid and grids
combined. (Sample sizes, and coefficients
of determination are indicated. )
Regression
Grid equation N R2
1608 Y = 8. 8 - 0. 57 (X) 7 0.61
1509 Y = 9.4 - 0. 63 (X) 8 0.75
1507 Y = 9.6 - 0. 66 (X) 7 0.71
1606 Y = 6. 7 - 0. 45 (X) 9 0.63
All Y = 1 1 - O - 0. 67 (X) 8 0.86


º
* * * * * * , a
* * * * * * * * * * *
- 2 = *
. . " " ' " ,w a
* * * * * * * **, * * *
." sº ºn a ºr "
sº * * * * * *
P - ºr- ºr -
Af
* * * * * * * * *.**
* * * * * * : * > . . ºf “
D º a
ſ
* * * ** ſ
e” a sº a
* * * *
p" sº
* * * *
* . . . . . .” ** *
is . . ºr a "º ºus ºr
3 tº a ra...ºr a tº ºr as
ſº º
* * * *S* º
* ºf
* * * * * * * * - **
*** * * * * * * * * º
“º “ sº e º 'º e º rººm ºs
f : * * *** * * * * * * * * * * **** -ºº:
- * wº
A
* : * ~ * * * * * * * * *
º tº- &
at ºf º B ** * º , , * * *
**** * * * * * * * *wº ºr fºr -rº, , ; *** * * * ºf:
- *** * **s g º a -
sº
* * * * * * * * * * *º -
* * * * * **** rºwº
* * * * * * * * * * * * tºº
* * * **, it tº wºn tº ºr
“tº ſº º ºf . ºw º
* * * * * ******* *
* * * * * * * *
º
º
* * * * * * ****** -
* * * * * * * * *sº wº
º * * * # * sº...? ºr rº Nº.
*** { };
* --
fºr . * sº
sº
ºw
* * * * * * * * * *
# * -
** º º
* * * *, * tº tº ºwn º tº
*** * : * r * * * **** * * º,
... º. A ºr tº “i º 1 v - 1 * * **, ºf
a wº a twº tºº.4 ºf , ºr º * * *
iſ tº a vºw nº sº.
a “t ºf tº 4 º ºvº.
A ºf ºf