~......;...' m’ VT“...W v? '
...,w..,. .5 .

.
:“

 

:- 0‘s
“u
v..l\
'. a

x “a.
t. .
Furl. r
EC» :1:
v I -
37;.“ “$4.33

 

I?-I~ ’ .JA .1..:... - V‘
«in ' ' ‘- ' ' ELIE-3 ...,
"I'LL" r .rj‘."' s. i v

A 4 u.‘
:7:

ma. ..
1 ”rm-R.
-u '

. ~ 31;:

1.62.99: “4;.“ '5'
-. lug". mdg", . M”
“u, 13319? P.

:33" 'n . arm <
.‘qug'u
, m

.‘h I 'u- ‘ r
gnu «mm

x.

Ml

g .

.53..
. :{v ,5.

 

 

 

 

933,???

‘ w - an
"an”.
. .73»,
'L't xi"‘.':
9':

Il' .J-u‘

nu.

h

ufifit

 

.
.l _ -
' ‘ . ., '1 ~ ’ Q: r» " 'J
' ' 'A ‘ ” " ' . " - ' , ' ~ a ." ' . , 2" ..f‘ .‘I.’ 1
F 3 . ‘ . ,4' >3! ‘ r)_. .7 ‘. .r A R v . '_ _ ”a" magnarw; k:
. , . H ., . y... , _ , ,M’ q '.j Fr". _,
- . ’ "f a ‘ a . '
5' ‘ . , ' : 'l‘ ,; h - 1 fl . . v ’ ' ' " _ ' .. . .. .’,r .' 911
'. ‘ ‘ ". _ r _ ., . ‘ f. 7‘ ..’ ‘fi; ‘ , ' ., “I . l'!‘c""""""
. ‘ . U? VI

' 7'
”mm".
n. .. .,... .

,, . "r‘mt

 

 

'l
. ‘ , ‘,' ' ' . , n. - ' .2. ”5... a. -,
— . . , , v. ‘ ‘ .' '- J. ,. u. ,. ‘ . .‘ ,.~ ., .. ,‘ , , 3-. ..‘
‘ , . ' ’ - A - - . . ...‘. .34. 173W" . ' "': g; tn-«W‘ . , , , é " .., ;;*,;,_',:,.;,-::
* . - v ~~ . ' N ' ' r , , r . 7 . “J... .,.J;’p-u_ a (o ’ . ~.- u: .— n" n- ,n, m.-. U..-
’ r i ' - ' ‘ ‘ I 1 v r . “a "in": a"?! ran , ».. tun-w. ‘- . - rvv — («L- .-
‘ ‘ I . I a "J . ' v u r: g , . q a "pv‘fv’g {x Y. 45‘ u a .3... g .
. a ‘ n ‘ mm m ' . . M , . * .- v . ' m ,2, . v 5 "'3.
:q. v :9; ». X1. . .70 .- 1‘1, 1V.-." 71’», v- Q, anw w.

. up
.,».. - . 35 - I ,,. . 3., . . . ”v.3... .5:
q n ‘ rm.’ ' ' A V ‘ , , ,. y .- . .1" .,.,, ”a“-
. . 0.... ...... 3”,. H",
. ‘ . ~41
-( ‘ ; fl . w.
ydmgvv.‘ .. . r , ,-
—.-»y.,»“. ..-v.,. , "H”... ”r . . . . .
‘ . .,..A, . .:.,-., .,.. ,, , w, ..
. . ,4 '. , ,. . . ‘- , ... V, ... , ”flaw
e. . - ,V « - .nu t rtmv‘vm
.. , .q. u. n . ”w .
_.
‘4’»

 

a
:u
.

‘ .
3.45
'l‘ I
a. .

 

,. , n.-
s ., . . Wynn..." m-
.E. . r‘ .. r‘.""¢!~" M~,r.—.

MiCHlGAN STATE LIBRARIES

1 111m! I’Hllllll’llllIll/ll!!!ll

I
3 1293 00910 1923

ll

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the
dissertation entitled

SIZE RELATED HABITAT USE IN A RIFFLE DWELLING FISH:
THE LONGNOSE DACE (RHINICHTHYS CATARACTAE)

 

presented by

DENNI S MICHAEL MULLEN

has been accepted towards fulfillment
of the requirements for

Ph . D . d . ZOOLOGY
egree 1n

 

 

a

Major professor

Date 5//¢9/;/
/ I

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

-.._—-—- _—

 

LIBRARY
Michigan State
; University

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE
11" a ‘ L231}
3 .5 :5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
cmmmi

SIZE RELATED HABITAT USE IN A RIFFLE DWELLING FISH: THE
LONGNOSE DACE (RHINICHTHYS CATARACTAE)

BY

Dennis Michael Mullen

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Zoology

1991

ABSTRACT

SIZE RELATED HABITAT USE IN A RIFFLE DWELLING FISH: THE
LONGNOSE DACE (RHINICHTHYS CATARACTAE)

BY

Dennis Michael Mullen

Most fish in mid-order streams exhibit across-habitat
size specific habitat segregation. Usually, smaller fish
are found in shallow areas, while larger fish are found in
pools. While several studies have examined the factors that
cause this pattern, little attention has been paid to the
patterns of size specific habitat use of fish within one
habitat type. The purpose of this study was to examine the
extent and cause(s) of size specific habitat use (within one
habitat type) in a common riffle dwelling fish: the longnose
dace (Rhinichthys cataractae).

A survey of the habitat use patterns of juvenile and
adult dace (in the Ford river system in Dickinson County,
Michigan), using kick seining and electrofishing, indicated
that adult dace use faster, deeper areas of riffles and
larger substrates than juveniles (< 67 mm Tl). Habitat
choice experiments conducted in a multiple habitat maze
indicated that dace > 55 mm Tl prefer faster areas in the
absence of other fish, and dace < 55 mm Tl prefer slower
areas. A laboratory flow chamber, used to determine maximum
swimming velocities of dace, indicated that dace < 55 mm T1
were unable to swim faster than 40 cm/sec and were,

therefore, excluded from faster areas of riffles. Instream

cages, with only one shelter, were used to examine the
effect of adult dace on the shelter use of juveniles (> 55
mm Tl). With adults present, juveniles used the shelter
significantly less than they did when the adults were
absent.

Manipulations of adult densities in enclosed riffles
were used to examine the effect of adults on the habitat use
of juveniles, over the first two years of their life.
Juvenile habitat use was not affected by adults densities
until they reached a size of 60 mm Tl. After that size, the
juveniles responded to reductions in adult densities by
moving into the faster areas. Adult density manipulations
had no effect on the depth or substrate use of the
juveniles.

The pattern of size specific velocity segregation
exhibited by longnose dace appears to result from a
combination of physical constraints and intra-specific

competition, acting sequentially as longnose dace grow.

To Susan; my inspiration when I needed it most

iv

ACKNOWLEDGMENTS

My appreciation is extended to the members of my
advisory committee; Dr. Thomas Burton, Dr. William Cooper,
Dr. Donald Hall and Dr. William Taylor, for their, patience
and advice over the course of this study. Dr. Thomas
Burton, Sue Eggert and Steve Mero deserve special thanks for
their advice, invaluable assistance and patience during the
field portions of this study. I would also like to thank
the following for their field assistance and advice; Dr.
Jean Stout, Deborah Repert, Bob Stelzer, Mike O'Malley,
Steve Marod and Natalie Briethaupt.

My parents; Mike and Shirley Mullen, have been my major
source of support and motivation throughout my career.
Without their love, none of this would be possible.

This rersearch was funded by; An American Museum of
Natural History: Theodore Roosevelt Grant and a Michigan
Polar-Equator Club Research Grant. Logistical support was
provided by a U.S. Navy grant to Dr. Thomas Burton.

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES
INTRODUCTION
STUDY SITE
CHAPTER ONE: HABITAT USE SURVEY
INTRODUCTION
METHODS
HABITAT SEGREGATION
HABITAT SELECTION
DIET ANALYSIS
RESULTS
HABITAT SEGREGATION
HABITAT SELECTION
DIET ANALYSIS
DISCUSSION
CHAPTER TWO: BEVAVIORAL STUDIES
INTRODUCTION
METHODS
HABITAT CHOICE EXPERIMENTS

SHELTER COMPETITION STUDY
MAXIMUM SWIMMING VELOCITY

vi

11
12
13
13
18
19
20
20
26
29
33
40
41
42
42

45
48

TABLE OF CONTENTS CONT.

RESULTS
HABITAT CHOICE EXPERIMENTS
SHELTER COMPETITION STUDY
MAXIMUM SWIMMING VELOCITY
DISCUSSION
CHAPTER THREE: RIFFLE MANIPULATIONS
INTRODUCTION
METHODS
RIFFLE MANIPULATIONS
DIET ANALYSIS
INVERTEBRATE DISTRIBUTION ON SUBSTRATES
RESULTS
RIFFLE MANIPULATIONS
DIET ANALYSIS
INVERTEBRATE DISTRIBUTION ON SUBSTRATES
DISCUSSION
CHAPTER FOUR: MORPHOMETRIC ANALYSIS
INTRODUCTION
METHODS
RESULTS
DISCUSSION
SYNOPSIS

LITERATURE CITED

vii

60
61
61
62
66
67
69
69
99
101
104
115
116
116
119
119
123

129

LIST OF TABLES

Table 1: Chi-square contingency table analysis

of the first order interactions involving size

from the 1988 riffle survey. Data are the

actual numbers captured in each cell with the
chi-square values for each cell and the sign of

the difference (observed - expected) N = 2994

longnose dace. 24

Table 2: Chi-square contingency table analysis

for associations among habitat variables for

Juveniles (N = 783), adults (N = 857) and a

random sample of the available habitat (N = 450)

for 1988. Data are the actual numbers captured

in each cell with the chi-square values for each

cell and the sign of the difference (observed -
expected). 25

Table 3: Dietary overlap between juvenile

(n = 35) and adult (n = 45) longnose dace
captured in the first week of September 1988.
Overlap is calculated according to Schoener
(1970).

31

Table 4: Habitat specific densities of selected
insect families in the Ford river on September
10 1988. Student t-tests were used to test for
significant differences between the habitats. 32

Table 5: Predictions of 5 possible explanations

for size specific habitat segregation in
longnose dace. See text for more details. 37

viii

gage
Table 6: Juvenile and adult densities
(numbers/m2) in each treatment during each of
the riffle manipulation studies in 1989 and
1990. The densities are corrected for
estimated capture efficiencies of 50 % for
juveniles and 60 % for adults (see chapter 1
for details on the determination of capture
efficiencies). The values in parentheses with
the 1989 data indicate the pre-manipulation
densities for each treatment. Pre-manipulation
densities for the 1990 studies are the control
densities. 71

Table 7: Comparison of the habitat

availabilities of each habitat variable in

each of the three riffles used in the 1989

riffle manipulation studies. The data are

the proportions of each habitat variable in

each riffle. 72

Table 8: Contingency table analysis (using

chi-square goodness of fit) of first order

interactions between size and each of the

habitat variables in unmanipulated (Initial)

and manipulated riffles in the June 1989 riffle
manipulation study. The sign next to

chi-square values greater than 0.5 indicates

the direction of the difference (observed -

expected). 74

Table 9: Contingency table analysis (using

chi-square goodness of fit) of first order

interactions between the control 1 and low

adult treatments and the use of each of the

habitat variables by the juveniles in each

treatment for each riffle manipulation study

in 1989. The sign next to chi-square values

greater than 0.5 indicates the direction of

the difference (observed - expected). 76

Table 10: Contingency table analysis (using
chi-square goodness of fit) of first order

interactions between treatments and the use

of each of the habitat variables by the

juveniles in each treatment for each riffle
manipulation study in 1989 and 1990. The sign

next to chi-square values greater than 0.5

indicates the direction of the difference

(observed - expected). 77

ix

Table 11: Contingency table analysis (using
chi-square goodness of fit) of first order

interactions between size and each of the

habitat variables in unmanipulated (Initial)

and manipulated riffles in the August 1989

riffle manipulation study. The sign next to

chi-square values greater than 0.5 indicates

the direction of the difference (observed -

expected). 79

Table 12: Contingency table analysis (using
chi-square goodness of fit) of first order

interactions between size and each of the

habitat variables in unmanipulated riffles at

the start of each riffle manipulation study

in 1990. The sign next to chi-square values

greater than 0.5 indicates the direction of

the difference (observed - expected). 82

Table 13: Contingency table analysis (using
chi-square goodness of fit) of first order

interactions between juvenile habitat use in

the low adult riffle before manipulation of

adult densities in August 1989 and juvenile

habitat use in that riffle after adult removal.

The sign next to chi-square values greater than

0.5 indicates the direction of the difference

(observed - expected). 98

Table 14: Dietary overlaps (Schoener 1970)
between adult and juvenile longnose dace
captured in unmanipulated riffles at the start
of each riffle manipulation study in 1989 and
1990. B = Baetis, C = Chironomidae, G =
Glossosoma, H = Hydropsychidae, Hp =
Heptageniidae, L = Leucotrichia, P = Psychomyia,
R = Rhyacophilidae and S = Simuliidae.
Subscripts refer to size classes based on head
capsule lengths where; S = 0 - 0.5 mm,

M = 0.5 - 1.0 mm and L = > 1.0 mm. Values in
parantheses indicate the percentage of the food
items in the diets of the juvenile and adult dace. 100

gage
Table 15: Dietary overlaps (Schoener 1970)

between juvenile longnose dace captured in low
adult density and control treatments during the
riffle manipulation studies in 1989 and 1990.

B = gagtis, C = Chironomidae, H =
Hydropsychidae, L = Leucotrichia, P =

Psychomyia and S = Simuliidae. Subscripts refer
to size classes based on head capsule lengths
where S = 0 ~ 0.5 mm and M = 0.5 - 1.0 mm.
Values in parentheses indicate the percentage

of the food items in the diets of the juvenile
dace in each treatment. 102

Table 16: Results of Kruskall-Wallis

comparisons of invertebrate densities on each

of three exposure surfaces (N = not exposed,

P = partially exposed and F = fully exposed). 103

Table 17: Results of principal component

analysis on eleven shape measurements from

87 longnose dace. Factors with loadings

greater than 0.2 (on a scale of 0 - 1)

were considered major factors for a component.

Factors are listed in order of decreasing

loadings. The shape variables designated by

each letter are illustrated in Figure 24. 120

xi

LIST OF FIGURES

Figure 1: Study site on the Ford River system

in Dickinson County, Michigan USA. The

darkened stretches indicate the specific

locations of the riffles used in this study. 9

Figure 2: Size frequency distributions of
longnose dace capturedin fast (N = 25) and
medium (N = 16) velocity areas in June 1987. 21

Figure 3: Size frequency distributions of
longnose dace captured in fast (N = 32) and
medium (N = 46) velocity areas in August 1987. 22

Figure 4: Chi-square analysis of the depth

distribution of juvenile (N = 485) and adult

(N = 539) longnose dace captured in 1988.

Positive values signify selection of that

depth cell while negative values signify

avoidance of that depth cell. 27

Figure 5: Chi-square analysis of velocity
distributions of juvenile (N = 485) and
adult (N = 539) longnose dace captured in 1988. 28

Figure 6: Chi-square analysis of substrate
use patterns of juvenile (N = 485) and adult
(N = 539) longnose dace captured in 1988. 30

Figure 7: Maze design for the habitat choice
experiments. The maze was 0.5 m deep and

situated in a riffle so that the desired

velocities were achieved in each channel. 43

Figure 8: Mean time spent by all size classes

in each of the three habitats of the maze.

Bars indicate standard errors. The Kruskal-

Wallis (non-parametric analysis of variance)

results were significant for all size classes

and the Tukey-type multiple comparisons indicate

that use of the fast habitat was significantly
different than use of the slow habitat for all

size classes. 52

xii

Figure 9: Mean time/use of shelters by juvenile
longnose dace; before adults were added (J1),
after adults were added (J+A) and after adults
were removed (J2) from the experimental cages.
Bars indicate standard errors. The Kruskal-
Wallis (non-parametric analysis of variance)
results were significant and the Dunnetts-type
test indicates that the shelter use of the
juveniles after the adults were added was
different than the shelter use before the
adults were added.

Figure 10: Maximum burst velocities of 17
longnose dace.

Figure 11: Design of riffle manipulations used
in 1989 on Twomile Creek.

Figure 12: Chi-square analysis of the depth
distribution of juvenile and adult longnose
dace captured in each treatment during the
June 1989 riffle manipulation study.

Figure 13: Chi-square analysis of the depth
distribution of juvenile and adult longnose
dace captured in each treatment during the
August 1989 riffle manipulation study. ‘

Figure 14: Chi-square analysis of the depth
distribution of juvenile and adult longnose
dace captured in each treatment during the
June 1990 riffle manipulation study.

Figure 15: Chi-square analysis of the depth
distribution of juvenile and adult longnose
dace captured in each treatment during the
August 1990 riffle manipulation study.

Figure 16: Chi-square analysis of substrate
use patterns of juvenile and adult longnose
dace captured in each treatment during the
June 1989 riffle manipulation study.

Figure 17: Chi-square analysis of substrate use
patterns of juvenile and adult longnose dace
captured in each treatment during the August
1989 riffle manipulation study.

xiii

54

55

63

75

80

83

85

86

87

Figure 18: Chi-square analysis of substrate use
patterns of juvenile and adult longnose dace
captured in each treatment during the June 1990
riffle manipulation study.

Figure 19: Chi-square analysis of substrate use
patterns of juvenile and adult longnose dace
captured in each treatment during the August
1990 riffle manipulation study.

Figure 20: Chi-square analysis of velocity
distributions of juvenile and adult longnose
dace captured in each treatment during the June
1989 riffle manipulation study.

Figure 21: Chi-square analysis of velocity
distributions of juvenile and adult longnose
dace captured in each treatment during the
August 1989 riffle manipulation study.

Figure 22: Chi-square analysis of velocity
distributions of juvenile and adult longnose
dace captured in each treatment during the
June 1990 riffle manipulation study.

Figure 23: Chi-square analysis of velocity
distributions of juvenile and adult longnose
dace captured in each treatment during the
August 1990 riffle manipulation study.

Figure 24: Outline of an adult longnose dace

showing the landmarks and measurements used for
the analysis of shape change with growth.

xiv

89

9O

92

93

95

97

118

INTRODUCTION

All animals exhibit some degree of habitat selection
over the course of their lives. Many motile animals exhibit
specific patterns of habitat selection that occur on both
small scales (daily or seasonal patterns) and large scales
(ontogenetic patterns). Daily patterns of habitat selection
involve movements between feeding and resting/shelter
habitats (eg. diel offshore movements for foraging by the
golden shiner (Notomigonug crysoleucas) (Hall et. a1.
1979)), while seasonal patterns are usually associated with
reproduction (eg. seasonal waterfowl migrations).
Ontogenetic patterns are innate patterns of habitat
selection that are usually associated with age/size or
developmental stage (eg. aquatic and terrestrial phases in
the life cycles of many species of amphibians) (see Werner
and Gilliam (1984) for a review of ontogenetic niche
shifts).

The choice of a habitat and the timing of habitat
shifts can be affected by many factors, including; foraging
returns, predation risk, inter-specific competition, intra-
specific competition and physical constraints and/or

requirements. Foraging returns have been shown to affect

the habitat choice, and behavior within a habitat, of many
organisms. In the absence of predators, bluegill sunfish
(Lepomis macrochirus) of all sizes foraged according to the
predictions of the optimal foraging model, switching
habitats when one habitat became more profitable than the
other habitats (Werner et. al. 1983a). When predators
(largemouth bass: Micropterus salmoides) were added to the
system, however, small bluegills tended to restrict their
habitat use to the protected, vegetated areas near the
shores of the experimental pond (Werner et. al. 1983b).
Juvenile aquatic insects (Notonecta hoffmanni) altered their
pattern of habitat use in the presence of predatory adults
of the same species (Sih 1982). When predatory adults were
present, the small juveniles used the food rich habitat
(pool center) less than they did when the adults were
absent. Larger juveniles remained in the center of the
pools with the adults present, but altered their behavior by
reducing the amount of movement and, therefore, the amount
of prey captured, while foraging.

Habitat choice may also be a function of the physical
requirements of the individual or species. Many species of
fish are stenothermal and only exist in areas that fall
within the range of thermal tolerance exhibited by that
species. This may be one of the factors behind non-
reproductive seasonal migrations in stream fishes. Brook

trout (Salvelinug fontinalis) in two Ontario streams moved

3
upstream as the temperatures in the downstream areas reached
249 C (near the lethal temperature of the species) (Meisner
1990).

Inter-specific and intra-specific competition may also
affect habitat use. In the face of competition with riffle
sculpin (Cottus gulggug), speckled dace (Rhinichthys
osculus) forages in stream margins, but, when sculpin are
not present, due to high temperatures, the dace utilize the
riffle habitats (Baltz et. al. 1982). Territorial juvenile
coho salmon (Oncorhvnchus kisutch) actively chase other
juveniles of the same species out of their feeding
territories in stream riffles (Puckett and Dill 1985). When
territories are limiting, the juveniles that are not able to
hold a territory adopt alternate strategies. These fish
either move into pools to forage, or exist in the riffles by
moving between the territories of the dominant juveniles.

In either case, intra-specific competition is forcing some
members of the population to utilize a different habitat or
to alter their behavior within a habitat.

A common result of predation risk, inter-specific and
intra—specific competition is size related habitat
segregation. Many species of stream fishes segregate the
local habitat according to size, with smaller fish occurring
in areas not occupied by larger individuals of the same
species (Mahon and Portt 1985, Power 1984, 1987 and

Schlosser 1987). In most cases, the larger species, or

4
individuals, occur in the pools, while the smaller species
or individuals occur in the shallow areas or riffles.
Smaller fish either, are not able to compete with the larger
fish for space in the pools or, use the shallow areas as a
refuge from the predators found in the pools. In some
cases, the smaller fish will use the pools when the larger
fish are absent (Power and Mathews 1983). Larger fish, on
the other hand, may avoid the shallow areas due to a high
predation risk from avian or terrestrial predators (Power
1987).

Size related habitat segregation may also be produced
by size related energetics. Smaller fish may not be able to
cope with the rigors of the adult habitat (especially in the
case of riffle dwelling fishes). As the profitability of a
habitat is measured in terms of energy gained per energy
expended while foraging, the profitability of a habitat is
as much a function of the physical nature of the habitat as
it is a function of prey quality and availability. The
ability of an organism to utilize a habitat may be strongly
influenced by physical factors such as temperature and
current.

Many species of stream fishes exhibit morphological
adaptations to the riffle habitat (Hynes 1970). Bottom
dwelling riffle fish tend to be dorsal-ventrally flattened
with subterminal mouths for picking food off of the bottom.

The swimbladders of riffle dwelling fish may be reduced in

5

size (Gee 1968). Most species of riffle dwelling fish start
life in the stream margins and undergo an ontogenetic shift
into the riffle habitat as they grow. These fish may also
undergo a morphometric (shape) change that adapts them to
the riffle habitat as they grow. If so, the timing and rate
of shape change may affect the ability of these fish to use
the riffle habitat.

Size related habitat segregation by stream fishes is a
fairly ubiquitous phenomenon in mid-sized streams throughout
North America (Mahon and Portt 1985), and has been well
studied (Power 1984, 1987, Power et. al. 1985, Schlosser
1987, 1988a and 1988b). Little attention, however has been
paid to the patterns of size related habitat use within a
specific habitat type, especially within the riffle habitat.
Since the riffles serve as a refuge for the smaller
individuals of pool dwelling species, as well as the
permanent habitat of all sizes of riffle dwelling species,
there is a potential for strong inter and intra-specific
interactions within the riffle habitat. The purpose of the
studies reported here was to examine the nature, degree and
cause(s) of size related habitat segregation within the
riffle habitat. Specifically, this study examined the
pattern of size related habitat use by the longnose dace
(Cyprinidae: Rhinichthvs cataractae Valenciennes).

The longnose dace is a ubiquitous stream minnow that

utilizes quiet stream margins for the first few weeks of

6
life and moves into stream riffles as it grows. Adult
longnose dace spawn in riffles during early summer when the
daily maximum temperature exceeds 15°<C (Bartnik 1970). The
fry emerge during mid-summer and recruit into the stream
margins. They remain there for several weeks and move into
the riffles in mid to late August. There is some evidence
to suggest that the timing of the habitat shift into the
riffles is affected by competition with blacknose dace
(Rhinichthys atratulus) fry (Gibbons and Gee 1972, Gee
1972). After moving into the riffle habitat, the longnose
dace tend to move into areas of faster water as they grow
(Gibbons and Gee 1972). The result is size related
partitioning of the riffle habitat by current velocity.

As longnose dace grow, the relative size of their
swimbladder decreases as does their ability to adjust the
swimbladder size to alter buoyancy (Gee 1972). The ability
to adjust their swimbladder size may allow the fry and
juvenile longnose dace some plasticity in habitat use that
is not experienced by the adults. Adult longnose dace may
be more morphologically adapted to the riffle habitat than
are the fry and juvenile longnose dace. Although size
related morphometric differences (other than swimbladder
volume) in longnose dace have not been demonstrated,
different populations of longnose dace in Nebraska streams
vary significantly in: caudal peduncle depth, pre-dorsal

length and snout length (Woodman 1986), all of which may

7
affect the ability of a longnose dace to utilize the riffle
habitat. Morphometric differences between the size classes
of longnose dace may contribute to size related habitat
segregation.

Longnose dace feed primarily on benthic insects of the
orders: Diptera (Chironomidae and Simuliidae), Ephemeroptera
(Baetidae and Heptageniidae) and Tricoptera (Hydropsychidae)
(Reed 1959, Gerald 1966, Pappantoniou and Dale 1982). The
proportions of these food items in the guts of longnose dace
change as the longnose dace grow. Chironomidae make up the
major portion of the diets of small longnose dace while
Simulidae and Hydropsychidae are more important in the diets
of larger longnose dace (Gerald 1966). Although longnose
dace were observed foraging in the early morning hours in
this study, longnose dace studied in Canadian rivers foraged
exclusively at night (Culp 1989, Beers and Culp 1989) (even
though most species of stream minnows are more active during
the crepuscular periods (Helfman 1986)). However, the
longnose dace that they studied rested behind shelters
during the daytime in the same areas that they use for
foraging (Culp 1989).

The objectives of this study were: to determine the
nature and degree of size related habitat use by longnose
dace (chapter one), to conduct behavioral studies that
investigate possible causes of size related habitat use

(intra-specific competition and physical constraints)

8
(chapter two), to conduct whole riffle manipulations of
longnose dace densities to verify the results of the
behaviotal studies (Chapter three), and to examine the
possibility that size related shape differences contribute

to size specific habitat segregation.

Study site:

This study was conducted in the Ford River system in
Michigan's upper Peninsula. All study riffles were located
on the Ford River and Twomile Creek (the major tributary to
the Ford) in Dickinson County, Michigan at T43N, R30W
(46908' north latitude and 87°54' west longitude), between
the towns of Channing and Ralph (Figure 1). The Ford and
Twomile are third order streams above their confluence, and
the Ford is a fourth order stream below the confluence.
Studies were conducted in riffles located in three areas of
the Ford and Twomile (Figure 1). These riffles were all
cobble/boulder riffles and ranged in depth (during normal
summer flows) from 5 cm to 35 cm. Widths varied from 8 to
13 m in the Ford River, and from 5 to 10 m in Twomile Creek.
Velocities reached a maximum of 120 cm/sec in the study
riffles. The Ford River is considered a marginal trout
river with maximum summer temperatures exceeding 24° C,
while Twomile Creek remains a few degrees cooler than the
Ford. The surrounding watershed is primarily glacial till

with granite bedrock and a few shale outcroppings. The

ca.

 

.mosum was» cfl.omms mmamuflu man no mcofluoooH ofluflowmm
05 mumoaofi monoumuum 605033 05. .«mo 3623:
.>UCSOO somcflxufio Ca Emummm Hm>fim ouom may :0 mpflm hosum ”a ousmfim

5m 2 $9222 .3500 COmEon “Congo. 6:5 mm .2

 

Exor

52m Rom

 

9::ch
_ Em: xosc. ESE.

xooeo
.882,

_ xwouo 2E 03...

10
bank vegetation is dominated by; speckled tag alder (Alnus
rggosg), balsam popular (Populus balsamifera), northern
white cedar (Thuja occidentalis) and white pine (gigug

strobus).

CHAPTER ONE

HABITAT USE SURVEY

11

12

INTRODUCTION

To date, most research on size specific habitat
segregation by stream fishes has focused on pool dwelling
species (Power 1984, Schlosser 1987, 1988a, 1988b). Thus,
little is known about the pattern of size specific habitat
use of riffle dwelling fishes. The purpose of this study
was to examine the pattern of size specific habitat use and
diet in a common riffle dwelling fish, the longnose dace
(Rhinichthys cataractae). Longnose dace fry move out into
the riffles around late July or early August, shortly after
emerging and, as they grow, they move into the faster areas
of the riffle (Gibbons and Gee 1972). They are a highly
adapted riffle fish with subterminal mouths, for feeding on
benthic invertebrates, and neutral or negative buoyancy,
achieved by the ability to decrease the size of their
swimbladder (Gee 1968).

A two year survey of habitat use patterns of both
juveniles (age 1 - 2 years) and adults (> 2 years) was used
to document the degree of size specific habitat segregation
in longnose dace. Comparisons were made between the depth,
water velocity, and substrate use, and the diets of both
size classes of dace and between the densities of the major

prey items in the slower and faster areas of the riffles.

13

METHODS

Habitat Segregation:

During the summer of 1987, the size distribution of
longnose dace in two habitat types (defined by water
velocity: medium velocity = 25 - 45 cm/sec and fast velocity
= > 45 cm/sec) was sampled twice, once during late spring
and once during mid summer. These velocities were chosen to
match the velocity categories of Gibbons and Gee (1972).
Areas to be sampled were selected the day before sampling by
locating sections of riffles with the desired velocity range
and recording the length, width and velocity range of each
section. Sampling of an area consisted of kick seining into
a 3.2 mm mesh hand held seine (2 m wide and 1.5 m deep).

The seine was held by two people while two other people
vigorously kicked the area 2 m immediately upstream of the
seine. All the captured fish were held in a bucket until
the entire area was sampled. The sampling crew started at
the downstream end of the sampling area and worked upstream
2 m at a time until the entire area was sampled. All
captured fish were identified, measured to total length (T1)
and released.

On June 12 (shortly after the longnose dace spawning
season), four medium velocity and three fast velocity areas
(2 - 3 m wide and 10 -15 m long) were sampled, and on August

12, five medium and five fast velocity areas were sampled.

14
A students' t-test was used to compare the mean size of
longnose dace captured in the medium velocity areas with the
mean size of longnose dace captured in the fast velocity
areas.

A more detailed approach was used in 1988 to document
the pattern of size specific habitat use in longnose dace.
The habitat use of juvenile and adult longnose dace in nine
individual riffle sections (25 - 30 m in length) was sampled
every three weeks for twelve weeks. The riffles were
grouped into sets of three and each set was sampled every
three weeks with a different set being sampled every week.
Sampling consisted of slowly electrofishing upstream with a
boat electroshocker (1 - 2 amps, 240 volts) and capturing
every dace encountered in the riffle. The location of
encounter for each dace was marked with a colored flag (to
identify size class) and three habitat variables were
measured at each of these locations. The depth, predominant
substrate type (using Cummin's (1962) classification) and
mean water column velocity were recorded for each dace
captured. Velocities were measured using a Pygmy-Gurley
current meter, and the mean water column velocity was
measured at 0.6 depth. Predominant substrate type was
visually estimated. Mean water column velocity was used
because it is more characteristic of the immediate area than
focal point velocity, which, for a benthic fish, can be 0

cm/sec in the fastest and slowest areas of the riffle. The

15
mean water column velocity is more characteristic of
imvertebrate densities (orth and Maughan 1983) and of the
velocities that may be encountered in the area while
foraging. All dace were measured (T1) and released back into
the riffle. The size frequency distribution of the
population was monitored weekly. These data were used to
track the growth of the juveniles and facilitate proper
identification of the size class of a captured dace.

On three occasions (independent of normal sampling), a
three pass removal method (Zippin 1958) was used to
determine the efficiency of the sampling method. The
population estimate for each size class from the removal
method was divided into the number of each size class caught
on the first pass to produce an efficiency estimate.

The sampling technique used in 1988 was slightly more
efficient for the adults (efficiency = 60 i 4.8%) than for
the juveniles (efficiency = 50 i 5.0%). Given the
structural complexity of a stream riffle and the difficulty
of netting shocked fish in a fast current, these
efficiencies are fairly high. The time it took to sample a
riffle depended on the water level and the density of dace
in the riffle and ranged from one to two hours. Since the
capture efficiencies were different for the two size
classes, the capture data for each size class was corrected
for its respective capture efficiency before statistical

analysis was conducted.

16

A four dimensional contingency table and chi squared
goodness of fit was used to test for independence of the
four variables; velocity, depth, substrate and dace size.
The velocity, depth, substrate type, and size of each dace
captured was fit into one of two categories for each
variable: velocity categories were: slow (<40 cm/sec) and
fast (> 40 cm/sec), depth categories were; shallow (<15 cm)
and deep (>15 cm), and substrate categories were; boulder
and others (over 95% of the others were cobble). The
dividing line between categories for a variable was chosen
to be the median of the range used by about 95% of the dace
captured. For instance, for the variable velocity,
approximately 90% of the dace were captured at velocities
less than 70 cm/sec and 98% were captured at velocities less
than 80 cm/sec. As 80 cm/sec falls close to the 95% mark, 0
- 80 cm/sec was chosen as the commonly used range and 40
cm/sec as the dividing line between slow and fast. In order
to test for specific interactions between size and each of
the habitat variables, the original four dimensional table
was broken down into three two dimensional tables, all
involving size as one of the variables.

Contingency table analysis was also used to examine the
possibility that either size class utilized a specific
combination of the habitat variables. Tests for
associations were conducted for the variable pairs; depth -

velocity, depth - substrate and velocity - substrate, using

17
data from all the dace captured in each size class and using
data generated from the riffle at large (see Habitat
Selection).

In this analysis, segregation was defined as a
significant chi-square value in the contingency tables
involving size and a habitat variable. Positive results
from these analyses would indicate that the distributions of
the adults and juveniles along a given habitat variable were
different than would have been expected if size were not
important, and therefore, segregation occurred by that
variable. Positive results would not indicate habitat
preferences for either size class, nor would they indicate
any effect of one size class on the habitat use of the
other.

All of the measurements of the habitat use of longnose
dace in this study were made between the hours of 0700 and
1100 hr. Even though most species of stream minnows are
most active during the crepuscular periods (Helfman 1986),
snorkling observations in the Ford River indictated that
longnose dace were active during this period. However,
longnose dace studied in Canadian rivers foraged exclusively
at night (Culp 1989, Beers and Culp 1989). In these
studies, the longnose dace spent the daylight period behind
or under shelters in the same areas that they used for
foraging. It is not clear why the longnose dace in this

study were more active in the morning hours than the

18
longnose dace studied by Beers and Culp, however, since
longnose dace rest in the same areas they forage in, the
results from this study should also apply to the longnose

dace populations studied by Beers and Culp.

Habitat Selection:

On seven of the twelve sampling days during the summer
of 1988, the availability of each of the three habitat
variables was measured using ten evenly spaced measurements
on each of 5 evenly spaced transects along the entire length
of each of the three riffles sampled that day (150
measurements total). On five of the twelve sampling dates in
1988, a combination of equipment failure and afternoon
thundershowers prevented habitat availability measurements.
Habitat availability measurements were taken at least twice
at each set of three riffles and three times in one set of
riffles.

These data were pooled and used to test the hypothesis
that all adult and juvenile longnose dace caught on these
seven days used the habitat (defined by the three habitat
variables) at random. Velocity was divided into six 10
cm/sec cells, depth into four 10 cm cells and substrate type
into two cells and the availability data were used to
calculate the expected catches of adults and juveniles in
each cell assuming random habitat use. The chi-square

values were used to determine the range of a habitat

19
variable that each size class avoided (utilized in a
proportion that was less than the proportion of that range
in the environment) and the range of each habitat variable
that each size class selected (utilized in a proportion that
was greater than the proportion of that range in the

environment).

Diet Analysis:

On the last sampling date in the summer of 1988, all
the longnose dace were kept for diet analysis. Thirty five
juveniles and 45 adults were preserved in 70% ETOH and
transported to the laboratory. At the laboratory the
anterior loop of the gut was removed from each dace and the
contents were identified to the lowest taxonomic level
practical (at least to family) and enumerated. The dietary
overlap (Schoener 1970) was computed using the proportions
of the most prominent items in the diets of both size
groups.

Two days prior to the last sampling date three
replicate invertebrate samples were taken from fast and
three from medium velocity areas of a riffle in order to
compare the densities between habitats for the major prey
items of both size classes. Samples were taken with a 0.11

2 modified Hess sampler, and sample locations were chosen

m
with the aid of a current meter to identify areas with mean

velocities between 50 and 60 cm/sec (fast) and between 30

20
and 40 cm/sec (medium). All samples were preserved in 70%
ETOH and sorted at the laboratory. All the invertebrates
were identified (at least to family) and counted from these
samples and statistical comparisons of densities were made

on the major prey items.

RESULTS

Habitat Segregation:

In both June and August of 1987 (Figures 2 and 3), the
fast velocity areas contained larger longnose dace than did
the medium velocity areas. In June, 16 dace (mean T1 = 62.1
mm) were captured in the medium velocity areas and 25 dace
(mean T1 = 75.1 mm) were captured in the fast velocity
areas. The null hypothesis: The mean size of longnose dace
captured in medium velocity areas was not different than the
mean size of longnose dace captured in fast velocity areas,
was rejected (t = 2.74, p < 0.01). In August, 46 dace (mean
T1 = 60.2 mm - including 18 fry that recruited to the
riffles in late July) were captured in the medium velocity
areas and 32 dace (mean T1 = 81.1 mm - including two fry)
were captured in fast velocity areas. Once again, the null
hypothesis of no size difference between the areas was
rejected (t = 5.21, p < 0.001). As this result may have
been due to the recruitment of longnose dace fry into the

medium velocity areas, and the purpose of this study was to

21

 

 

 

 

 

 

 

 

 

 

 

5.0
I:
MEDIUM
4.0“ l- N
FAST
6 3.0- —
2
[U
I)
O
LU
E; 20«
1.0-
0.0 l l l I 7 l 1 1
47 51 55 59 63
SIZE (MM TL)

Figure 2: Size frequency distributions of longnose dace
captured in fast (N = 25) and medium (N = 16)
velocity areas in June 1987.

22

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

so
50‘ MEDIUM
' .\\V
FVMST
40‘
>-
L)
2
LU
I) 30- , . ,
C3
UJ
m
IL
20- i
1
I
10— I
1
l
I
I
. l
00 I I I BE
- TI I IITITIIT IIT’TIIT IflITrrI
29 33 37 41 45 49 53 57 en 65 69 73 77 81 85 e9 93 92>101
SHHE(NMAIIJ

\Figure 3: Size frequency distributions of longnose dace
captured in fast (N = 32) and medium (N = 46)
velocity areas in August 1987.

23
examine the habitat use of only the juveniles and adults,
the fry were removed from the dataset and the t-test was
conducted with just the juvenile and adult data. Even
without the fry in the medium velocity areas, there was a
significant difference between the size distributions of
longnose dace in the medium and fast areas (t = 3.42, p <
0.001) .
In 1988, data on the depth, velocity and substrate use

of 1,640 longnose dace were collected over the summer.
The four dimensional contingency table indicated that size,
depth, velocity and substrate type were highly dependent on
each other (chi-square = 87.85, p < 0.001). The two
dimensional contingency table analysis of size versus each
habitat variable indicated that size segregation occured by
depth, velocity and substrate type (Table 1 - a significant
chi-square value indicates that the two variables [i.e. size
and velocity] were not independent). Juveniles were found
more frequently in shallow, slow non-boulder areas and
adults were found more frequently in deep, fast boulder
areas than would be expected if size were not important.

Tests for associations between habitat variables for
juveniles, adults and the riffle at large (Table 2)
indicated that substrate type was not associated with depth
or velocity but that depth was associated with velocity for
each size class. Therefore, the depth at which both

juvenile and adult longnose dace were captured was a

24

Table 1: Chi-square contingency table analysis of the first
order interactions involving size from the 1988 riffle survey.
Data are the actual numbers captured in each cell with the chi-
square values for each cell and the sign of the difference
(observed - expected). N = 2994 longnose dace.

 

 

Size Depth Velocity Substrate type
Shallow Deep Slow Fast Boulder Other
Juveniles
Observed 800 766 1178 388 254 1312
chi-square 1.26+ 1.21- 4.80+ 11.50— 8.34- 2.01+
Adults
Observed 670 758 935 493 328 1100
chi-square 1.38- 1.34+ 5.26— 12.6l+ 9.15+ 2.21-

Significance p < 0.05 p < 0.001 p < 0.001

 

25

Table 2: Chi-square contingency table analysis for associations
among habitat variables for Juveniles (N = 783), adults (N = 857)
and a random sample of the available habitat (N = 450) for 1988.
Data are the actual numbers captured in each cell with the chi-
square values for each cell and the sign of the difference
(observed - expected).

 

 

JUVENILES ADULTS AVAILABLE
DEPTH X SUBSTRATE
Shallow Deep Shallow Deep Shallow Deep
Boulders
Observed 58 69 81 116 33 37
chi-square 0.73- 1.02+ 1.41- l.24+ 0.47 0.54+
Others
Observed 342 314 321 339 206 174
Chi-square 0.14 0.15 0.42 0.37 0.09 0.10
Significance NS NS NS
DEPTH X VELOCITY
Shallow Deep Shallow Deep Shallow Deep
Slow
Observed 332 257 288 273 210 166
Chi-square 3.21+ 3.36- 2.34+ 2.07- 0.53+ 0.60-
Fast
Observed 68 126 114 182 29 4S
Chi-square 9.76- 10.l9+ 4.43- 3.91+ 2.70- 3.06+
Significance p < 0.00 1 p < 0.001 p < 0.01
VELOCITY X SUBSTRATE
Slow Fast Slow Fast Slow Fast
Boulders
Observed 100 27 133 64 55 15
Chi-square 0.21 0.64- 0.12 0.24 0.21 1.07-
Others
Observed 489 167 428 232 321 59
Chi-square 0.04 0.12 0.04 0.07 0.04 0.20

Significance

 

26
function of the velocity at each capture location, perhaps
indicating that the juveniles and adults were utilizing a
specific combination of depth and velocity. However, the
pattern of the association (+'s and -'s in Table 2) between
depth and velocity was similar for both size classes, and
this pattern was similar to the pattern of association of
these two variables in the riffle at large ("available" in
Table 2). Therefore, the observed association in the depth
and velocity use of both size classes probably resulted from

the correlation of these two variables in the environment.

Habitat Selection:

Data on habitat selection was collected on 485
juveniles and 539 adults. Longnose dace of both size
classes selected depths between 10 and 19 cm (Figure 4).
Juveniles avoided depths over 20 cm and used depths less
than 10 cm in proportion to their availability while the
adults avoided depths less than 10 cm and deeper than 30 cm.
In both cases, the overall chi-square value was significant
(p < 0.001) indicating that neither size class utilized the
habitat (defined by depth) at random. Both size classes
showed a similar pattern of velocity selection (Figure 5),
avoidance of slow velocities (less than 10 cm/sec) and
selection of faster velocities (greater than 40 cm/sec).
Although selectivity was much stronger (the magnitude of the

chi-square values indicates the degree to which each size

27

 

 

§§f‘?=\

  
 
  

JUVENILES P < 0.001

D

ADULTS P < 0.001

 

 

CHI-SQUARE VALUE

 

 

 

-10—1

 

 

 

 

0-9 10-19 I 20-29 >30
DEPTH CELL (CM)

Figure 4: Chi-square analysis of the depth distribution of
juvenile (N = 485) and adult (N = 539) longnose
dace captured in 1988. Positive values signify
selection of that depth cell while negative values
signify avoidance of that depth cell.

28

 

JUVENILES P < 0.001
40- I:

ADULTS P < 0.001

 

 

 

CHI-SQUARE VALUE

 

 

 

 

 

 

I T f
0-9 10-19 20-29 30-39

T
40-49 >50

1

VELOCITY CELL (CM/SEC)

Figure 5: Chi-square analysis of velocity distributions of
juvenile (N = 485) and adult (N = 539) longnose

dace captured in 1988.

29
class used or avoided each velocity cell) for adults, the
overall chi-square value for both size classes was
significant (p < 0.001). The pattern of substrate selection
was also similar (Figure 6). Both size classes selected
boulders and avoided other substrates. Again, the overall
chi-square value for both size classes was significant (p <
0.05 for juveniles and p < 0.001 for adults), but the adults

showed a stronger degree of selectivity.

Diet analysis:

The diets of the 35 juvenile and 45 adult longnose dace
sampled in early September 1988 were similar (overlap =
83%), with slight differences occurring in the proportion of
Chironomidae, Baetidae, Simuliidae and Hydropsychidae in the
diet (Table 3). The juvenile diet consisted primarily of
Chironomidae, while the adult diet was slightly broader,
including more Baetidae, Simuliidae and Hydropsychidae.
These results are similar to those of Gerald (1966) who
studied the diets of 439 longnose dace from the Yellowstone
River, Montana, but differ markedly from those of
Pappantoniou and Dale (1982) who found that hydropsychids
dominated the diets of 52 longnose dace collected from the
Waccabuc River, New York. In general, these prey items
occured at higher densities in the fast areas than in the

medium velocity areas of the Ford River (Table 4).

30

 

 

.\\\\\\‘
JUVENILES P < 0.001

4
40 Cl

ADULTS P < 0.001

CHI-SQUARE VALUE
3

10‘

 

 

/2

 

 

0 x\\\\\

 

 

 

 

-10
BOULDER ' OTHER

SUBSTRATE TYPE ,

Figure 6: Chi-square analysis of substrate use patterns of
juvenile (N = 485) and adult (N = 539) longnose
dace captured in 1988.

31

Table 3: Dietary overlap between juvenile (n = 35) and adult (n
= 45) longnose dace captured in the first week of September 1988.
Overlap is calculated according to Schoener (1970).

 

 

 

Item Proportion

Juveniles Adults :le— Pa:
Chironomidae 0.86 0.75 0.11
Baetidae 0.02 0.08 0.06
Simuliidae 0.01 0.06 0.05
Hydropsychidae 0.02 0.06 0.04
Psychomyiidae 0.04 0.03 0.01
Others 0.07 0.00 0.07
Total :Pj - Pa: 0.34
Overlap 83%

 

32

Table 4: Habitat specific densities of selected insect families
in the Ford river on September 10 1988. Student t—tests were used
to test for significant differences between the habitats.

 

Item Density (#/m2) i 8.8. Significance

 

 

Medium Fast t P
Chironomidae 2775 i 55 4968 i 1168 -l.70 0.165
Baetidae 121 t 64 394 t 67 -2.93 0.043
Hydropsychidae 1358 i 531 5077 i 623 —4.54 0.011
Psychomyiidae 27 i S 3 i 3 1.19 0.39
Simuliidae 15 i 10 163 t 63 -2.32 0.081

 

33

DISCUSSION

Some degree of habitat segregation between adult and
juvenile longnose dace occurred by all three habitat
variables. Segregation by velocity was strongest (chi-
square = 34.17) with adults occurring in fast water in a
much higher proportion than expected while the juveniles
occured in the fast water in a much lower proportion than
did the adults (Table 1). Similarly, the adults utilized
boulders much more than expected and the juveniles utilized

the boulders in a lower proportion than the adults (chi-

square 21.71). Segregation by depth was the weakest (chi-
square = 5.19) and probably resulted from the positive
association between depth and velocity in the riffle at
large (Table 2). Since deeper waters were associated with
faster velocities, and the adults used the faster velocities
in a much higher proportion than did the juveniles, they
would also have used deeper waters in a higher proportion.
Despite the strong degree of habitat segregation, both
size classes selected similar velocity and depth ranges and
substrate types. In all cases the adults showed the
stronger selectivity, indicating that segregation resulted
from a higher proportion of the adults utilizing the fast

areas and boulders than the proportion of juveniles in those

areas, even though both size classes utilized those areas in

34

higher proportions than they occurred in the environment.

The actual cause of segregation is unknown, but,
several possible explanations exist. Adults may utilize the
faster areas of riffles and larger substrates as a refuge
from predators. Potential aquatic predators in the Ford
River include adult brook trout (Salvelinus fontinalis) and
burbot (Lota lota), both of which forage in riffles and
pools (personal observation). Differential predation rates
on the two size classes may be forcing the adults into the
refuge areas, resulting in the observed size segregation.
However, juvenile dace would also be susceptible to
predation by these predators. The mean length of the
juvenile size class exceeded 60 mm Tl by the end of June (at
least as large as the minnows used as bait by many trout
anglers). Segregation may arise from intraspecific
competition for the shelter provided by the faster areas
(where the cost of foraging should be high for a large
piscivore) and around boulders. By this scenario, the
juveniles and adults should show selection for faster areas
and larger substrates (assuming that these areas are not
completely limited). Alternatively, juveniles may lack the
physical ability to escape predation by moving into the
faster areas of the stream due to the strength of the
current. Under this scenario, the juveniles should show
avoidance of the faster areas but still show selection for

boulders as a refuge from predation.

35

Segregation may have also resulted from predation
pressure on adult longnose dace from terrestrial or avian
predators. Kingfishers, herons, mink and raccoons are all
potential longnose dace predators in the Ford River. In
general, these types of predators prey on larger fish (Power
1987 and Schlosser 1988a, 1988b) often confining these fish
to the deeper areas of the stream. Adult longnose dace may
be using the faster areas of the riffle because they are
positively associated with depth (although this is only
within a riffle and the depths are usually less than 25 cm
and may not offer much refuge) or more likely, because the
faster areas are areas of high turbulence which would
inhibit non-aquatic predators. Larger substrates may also
increase the turbulence and offer more hiding places from
these predators. Juvenile longnose dace, due to their small
size, should not be as susceptible to non-aquatic predators
and therefore should show less selectivity of velocity or
substrate types barring any other factors that might affect
juvenile habitat use.

Alternatively, adults may be utilizing the faster areas
of the riffles because they have higher invertebrate prey
densities. With the exception of Chironomidae, the major
prey items of longnose dace (Baetidae, Hydropsychidae and
Simuliidae) are more dense in the faster areas of the
riffles (Table 4, Orth and Maughan 1983 and Brown and Brown

1984). By this scenario, segregation should occur by

36
velocity but not by substrate size (invertebrate densities
are highest in areas dominated more by cobbles rather than
by boulders (Brown and Brown 1984). Juveniles should show
selection for faster velocities if competition for foraging
space in the fast areas explains segregation, or they should
show avoidance of faster velocities if physical constraints
on the juveniles explain segregation. There should be
selection for smaller substrates by both size classes.

Five explanations for habitat segregation have been
proposed and can be separated on the basis of the
predictions that each makes about segregation between the
size classes and selection by the juvenile size class (Table
5). According to these predictions, the results from this
study (segregation by velocity, depth and substrate type and
juvenile selection for faster velocities and larger
substrates) are consistant with the hypothesis that
segregation resulted from predation pressure on the dace by
aquatic piscivores and size class competition for the
shelter from predation that was offered by fast velocities
and large substrates. However, some other possible cause
of segregation may be acting on the habitat use of the dace.
For instance, this study used data collected over an entire
growing season and physical constraints on the juvenile dace
may be important early in the summer and not later in the
summer. Also, higher invertebrate prey densities may

combine with lowered predation risk to make the faster areas

37

Table 5: Predictions of 5 possible explanations for size

specific habitat segregation in longnose dace.

See text for more

 

 

 

details..
Predictions
. Segregation by; Juveniles select;
Explanation Velocity Depth Substrate Velocity Substrate
Predation
Aquatic
Physical
constraints yes yes no slow boulder
Competition yes yes yes fast boulder
Non aquatic yes yes yes neither neither
Invertebrate
density
Physical
constraints yes yes no slow others
Competition yes yes no fast others

 

38
more desirable to both size classes.
Most species of fish exhibit an optimal temperature for

growth. Young sockeye salmon (Oncorhynchus nerka) grow

 

fastest at about 15°C, but as the food ration is decreased
the optimal temperature for growth decreases (Brett 1971).
When surplus food is available, young sockeye reach their
highest growth rate at high temperatures (Biette and Geen
1980). The thermal preferences for brook trout (Salvelinus
fontinalis) increases from 8 - 11.5°C as fry to 17.5°C as
fingerlings (Peterson et. al. 1979). Longnose dace may move
into the faster areas of the riffle as they grow in response
to changing temperature optima, however the optimal
temperatures for growth of longnose dace of various sizes
are unknown. There is probably only a little variation in
temperature across a riffle and thermal preferences are
probably not the cause of size specific habitat use of
longnose dace. In addition, when food is abundant (as it is
in riffles), fish growth rates tend to be higher at higher
temperatures, which would occur in the slower, shallower
areas of the riffle and not in the deeper, faster areas
preferred by adult longnose dace.

This study relied on observational data and, although
it documents habitat segregation between the size classes,
it can only suggest possible causes for that segregation.
Experiments have been conducted to specifically test for the

effects of adult longnose dace on the habitat use and diet

39
of juvenile longnose dace, and to examine the ability of the
juveniles to physically utilize the faster areas. In
addition, an analysis of shape changes with growth was
conducted to examine the possibility that segregation
results from shape differences between the size classes.

These will be the subjects of the following chapters.

CHAPTER TWO

BEHAVIORAL STUDIES

40

41

INTRODUCTION

The pattern of velocity segregation between juvenile
and adult longnose dace detected in the habitat use survey
(chapter one) may result from size specific habitat
preferences. Juveniles may prefer slower habitats for
energetic reasons. The higher food densities in the faster
areas may not be enough to offset the energetic costs of
foraging in those areas. Since the ratio of surface area to
volume is higher for smaller fishes of the same shape as
their larger conspecifics, those costs should be
proportionally higher for smaller fish (Vogel 1988). For
practical purposes, the volume of a fish determines the
amount of muscle mass available and the surface area
represents the area that has to be moved against the
current. As the fish grow, they develop a larger muscle
mass for moving a proportionally smaller surface area
against the current. Smaller dace, therefore, may be
physically unable to utilize the faster areas of the riffle.

Interactive segregation may also be responsible for the
habitat segregation observed in the habitat use survey.
Adult and juvenile longnose dace may compete for shelters
from the current (or from predation) in the faster areas of
the riffle. If shelter space is not completely limited

there should be some overlap in habitat use with the

42
subordinate group using proportionally less of the preferred
habitat. This pattern was observed in the riffle survey
(Figures 5 and 6).

The purpose of the studies reported here was to examine
the size specific habitat preferences (in the absence of
other longnose dace) of various size classes of longnose
dace, to examine the effect of adults on the shelter use of
juvenile longnose dace, and to determine the maximum

swimming velocities of various sizes of longnose dace.

METHODS

Habitat Choice Experiments:

The habitat preferences of 18 fry (0+ years old: 37 -
51 mm T1), 18 juvenile (1+ years old; 55 - 64 mm T1) and 18
adult (2+ years old; 67 - 88 mm Tl) longnose dace were
tested one at a time in a maze designed to give fish a
choice of three habitats: slow (0 - 10 cm/sec), medium (25 -
35 cm/sec) and fast (40 - 50 cm/sec) current velocities
(Figure 7). The maze was constructed in a 1 mg, 50
centimeter deep cage built from treated plywood and 8 inch
hardware cloth and was situated in a riffle in the Ford
River in about 15 centimeters of water. Each habitat

contained substrates from the equivalent habitat in the

river. Longnose dace for testing were captured the day

43

FAST MEDIUM SLOW

40-50 cm/sec 25-35 cm/sec 0-1 0 cmlsec

 

ttn

 

 

0

release point
114 inch hardware cloth

..........................................................

 

 

 

 

Figure 7: Maze design for the habitat choice experiments.
The maze was 0.5 m deep and situated in a
riffle so that the desired velocities were
achieved in each channel.

44
before testing by kick seining in a nearby riffle and held
in small flow-through holding cages in a medium velocity
riffle (the holding cages contained rocks to provide cover
for the dace). Testing consisted of choosing a dace of the
appropriate size class (predetermined before each test) from
the holding cage, rubbing a small amount of fluorescent
powder (Cabela's Magic Powder) (the powder remained on the
fish for 15 - 20 min and facilitated observation in the
faster channel where turbulence was high) onto the dorsum of
the fish and placing the dace into the maze at the
downstream end of the slow channel. This location was
chosen because the dace quickly moved out of this spot,
usually exploring the slow channel before moving into the
medium or fast channels. Dace released in the medium or
fast channels usually found the nearest cover and remained
there for the duration of the test. The dace were allowed 5
minutes to adjust and then were observed from a spot behind
the cage for 15 minutes. The amount of time a dace spent in
a predetermined habitat was recorded. This gave the amount
of time spent in each habitat by longnose dace of each size
class. A dace was timed in only one habitat in order to
provide independence between habitats for statistical
analysis. All observations were made between 07:00 h and
10:00 h, and all the adults and juveniles were tested
between June 25 and July 5, 1987. The fry were tested

between August 20 and August 29, 1987. This was done in

45
order to provide data on as wide a range of longnose dace
sizes as possible (the fry had not recruited into the riffle
areas at the time of the June/July testing).

A Kruskal-Wallis (non-parametric analysis of variance)
test was used for each size class to test the null
hypothesis that there was no difference in the amount of
time spent in each habitat for that size class. When the
null hypothesis was rejected, a Tukey-type multiple
comparison of means (Zar 1984) was used to determine which

habitats were used differently.

Shelter Competition Study:

The effect of adults on the shelter use of fifteen
juvenile longnose dace was examined using a n3 (1 m long x
0.5 m wide) cages situated in a riffle in the Ford River.
Five cages were constructed of % inch treated plywood sides
with % inch hardware cloth screens at the front and back.
The bottom of the cages were filled with gravel and pebbles
from the river and the cages were situated so that a current
of 15 — 20 cm/sec was flowing through each cage. The cages
were placed in the river three weeks prior to use to allow
invertebrate colonization. One shelter was provided in each
cage. The shelter was a piece of clay tile, about 10 m wide
and 15 m long, that, when viewed from the side, was shaped
like an "L". The tile was laid in the middle of the cage so

that the short leg of the "L" was vertical to the substrate

46
and the long leg of the "L" was nearly horizontal to, and
suspended above, the substrate at the forward end.
Preliminary observations demonstrated that both sizes of
dace readily used the shelters and could easily move in and
out from under the shelter. A slow current was used in this
study because it was nearly impossible to observe the dace
in the cages when faster currents were used (the fluorescent
powder that was used for the observations of habitat
preference did not stay attached to the fish long enough to
be useful in this study). The forward screen produced
turbulence in the cages at higher velocities. Also, the
cages were left undisturbed for 24 h before each
observation, and over that time enough debris accumulated on
the forward screen to lower the velocities in the cages to
15 - 20 cm/sec. This debris was not removed until after
observations were made (removal of the debris disturbed the
cages and may have affected the results if this were done
before the observations were made).

Observations were made between 08:00 and 10:00 h over
three, three day periods between June 29 and August 11,
1990. The experiments were conducted as follows: 1)
Juvenile longnose dace (61 - 69 mm T1) were captured by kick
seining and placed in the cages (one to a cage) one day
prior to the initial observation, 2) The shelter use of each
juvenile was recorded during the initial observation and one

adult (75 - 90 mm T1) was placed in each cage immediately

47

after the observations were made, 3) The shelter use of the
juveniles in the presence of adults was recorded 24 hours
later and the adults were removed from the cages immediately
after the observations were made, 4) The shelter use of the
juveniles (without the adults) was recorded 24 hours later
and then all fish were released. This design was used to
control for the possibility that differences in shelter use
observed after the adults were added occurred because the
juveniles were becoming acclimated to the cages, and
therefore more active, instead of being due to the presence
of adults.

Prior to observation, each cage was approached slowly
from the downstream direction and the observer sat in a
chair behind the cage at a distance that allowed the
interior of the cage to be seen. The observer carefully
searched the cage for the dace and if the dace was not found
it was assumed to be under the shelter (this assumption was
verified at the end of the observation period). This was
the case for 42 of the 45 observations, in the other three
observations the dace was located outside the shelter. Once
the location of the dace was determined, the observer
recorded the total amount of time that the dace used the
shelter over a 15 minute period, and the number of times the
dace used the shelter. These values were used to determine
the average time/use for each juvenile longnose dace without

an adult present, with an adult present and again without an

48
adult present. In all cases, the adult, when present, spent
the entire observation period under the shelter. In nearly
all cases, when a juvenile exited the shelter it quickly re-
entered the shelter.

Kruskal-Wallis (nonparametric analysis of variance by
ranks) was used to test for an adult effect on the
time/shelter use of the juveniles. A Dunnett's-type test
for comparing a control group (initial observation) to each
of the other groups (with an adult present and the second
observation without adults) was used to examine the
possibility that a significant Kruskal-Wallis result was due
to an adult affect (control versus adults present) or due to
an acclimation affect (control versus the second adult free

observation) (Zar 1984, pg. 204 -205).

Maximum swimming velocity:

In order to examine the possibility that juvenile
longnose dace were excluded from faster habitats by virtue
of an inability to swim against the currents encountered in
the fast habitats, the maximum burst velocities of dace of
several sizes were measured. Measurements were made on 41
dace ranging in size from 41 to 103 mm Tl. Measurements
were made in a Clear plastic tube with a 4.9 cm inner
diameter that was screened off and attached at both ends to
flexible tubing (an input hose and an output hose). The

input hose was attached to a Teal model 2P390 pump capable

49
of pumping up to 2520 gph. A hand valve was placed on the
input hose to regulate the flow rate through the tube. The
whole apparatus (except the pump and valve) was submerged in
a Frigid Unit living stream.

The longnose dace were collected from the Ford River in
early May (1987) and transported to East Lansing in an
aerated ice chest. The dace were kept in a living stream
set at 10 °C (the temperature at which the dace were
collected), provided with shelter and fed four cubes of
tubifex worms every other day (the dace were observed eating
and appeared healthy throughout their captivity). Dace were
held for at least ten days before testing, and testing was
completed in four days. All dace were returned to the Ford
River after the completion of the study.

A dace was placed in the tube by removing the screen at
the outlet and replacing it after the dace was in the
apparatus. A strip of black electrical tape was wrapped
around the tube in the middle to give the dace something to
orient to (without the tape the dace tended to cower near
the outlet, with the tape the dace showed a strong tendency
to keep their heads within the area covered by the tape).

To start each trial the pump was turned on with the valve
closed and then the valve was opened slowly at a constant
rate (one complete turn per ten seconds). The dace

immediately oriented to the current and came to a resting

position on the bottom of the tube, normally with their

50
heads in the area covered by the tape. The valve was opened
at the constant rate until the dace was swept backwards in
the tube. At this point the velocity was held constant
until the fish made an effort to advance against the
current. This usually happened immediately after the dace
was swept out of position but in some cases the fish did not
attempt to swim until it contacted the outlet screen. If
the dace was not able to swim forward the trial was stopped
there. If the dace was able to advance and hold a position
the velocity was increased and the trial continued until the
dace could no longer advance (no trial lasted longer than
1.5 min.). At the end of the trial, the pump was turned off
(the valve was left at the setting at which the trial ended)
and the dace was removed from the apparatus. In order to
determine the final velocity, the apparatus was reassembled
and the pump turned on. The output from the apparatus was
collected in a bucket for 10 seconds and the volume was
measured. The volume was used to calculate discharge
(cmH/sec) which was divided by the cross sectional area of
the tube to determine velocity in cm/sec. This procedure
was repeated three times for each trial and the average
velocity was used. The standard errors from the three
velocity measurements ranged from 0.21 to 0.81 cm/sec with
an average of 0.43 cm/sec. The upper limit of the apparatus

was 60 cm/sec.

51

RESULTS

Habitat Choice Experiments:

In the habitat choice experiment, adult and juvenile
longnose dace used the fast habitat more than the other
habitats while the fry used the slow habitat more than the
others (Figure 8). Kruskal-Wallis (nonparametric analysis
of variance) tests for all size classes reject the null
hypothesis that the mean time spent in each habitat was the
same (fry; Hc== 6.73 P < 0.05, juveniles: Hc== 12.19 P <
0.005, adults; Hc== 6.23 P < 0.05). Tukey-type multiple
comparisons showed that the difference was due to the fast
versus slow comparisons for all size classes (for the large
size class, the fast versus slow comparison was not
significant at the 0.05 confidence level but was significant
at the 0.07 confidence level). Although the juveniles and
adults did not use the fast channel significantly more than

the medium channel, neither size class was limited by the

velocities in the fast channel.

Shelter Competition Study:

The fifteen juvenile longnose dace in this study spent
nearly the entire observation period under the shelter in
the absence of adults (in only one case did the juvenile
leave the shelter during the observation period). After

adults were added, however, the average time per shelter use

20

52

 

15:

10-

MEAN TIME (min)

 

Figure 8:

SLCWV
MEUUM
D I

FAST

 

 

 

 

 
 

 

 

 

 

 

\\\\\\\\\\\V—-—
\\\\\\\\\\\\\V—-—-

 

 

 

I

FRY ' JUVENILES ' ADULTS
TREATMENT

 

Mean time spent by all size classes in each of
the three habitats of the maze. Bars indicate
standard errors. The Kruskal-Wallis (non-
parametric analysis of variance) results were
significant for all size classes and the Tukey-
type multiple comparison indicate that use

of the fast habitat was significantly different
than use of the slow habitat for all size
classes.

53
dropped significantly (Figure 9). In seven cases the
juvenile shelter use was not affected by the presence of the
adults, but in three cases, the juveniles did not use the
shelter when the adults were present and in five cases the
juveniles exited and re-entered the shelter a number of
times (1 - 6 exits). In all cases the adults spent the
entire observation period under the shelters. After the
adults were removed, the juvenile shelter use returned to
the level observed before the adults were added (Figure 9),
thirteen of the juveniles did not leave the shelter during
the observation period while two exited and re-entered the
shelter once. The Kruskal-Wallis results were significant
0% = 11.14, P < 0.005). The Dunnett's-type test
contrasting the initial observations with the adult present
observations and with the second adult free observations
indicated that the difference detected in the Kruskal-Wallis

test was due to an adult affect (q = 2.27, P < 0.01) and not

due to an acclimation affect.

Maximum Swimming velocity:

Of the forty one longnose dace tested, twenty four,
ranging in size from 68 to 107 mm T1 were able to swim
faster than the 60 cm/sec limit of the apparatus. As a
result, burst velocity measurements were made on 17 longnose
dace ranging in size from 41 to 74 mm Tl (Figure 10).

Maximum burst velocity increased with size, with dace less

54

 

 

 

 

 

20
P < 0.01
I I
15* l
|
us
(I)
2
E
5: 10- -
.z
,<
u:
‘2
5..
O T I I
J1 J+A J2
TREATMENT
Figure 9: Mean time/use of shelters by juvenile longnose

dace: before adults were added (J1), after adults
were added (J+A) and after adults were removed
(J2) from the experimental cages. Bars indicate
standard errors. The Kruskal-Wallis (non-
parametric analysis of variance) results were
significant and the Dunnetts-type test indicates
that the shelter use of the juveniles after

the adults were added was different was different
than the shelter use before the adults were
added.

Maximum Swimming Velocity (cm/sec)

55

 

 

 

 

60
A A
A
A
A
A
50—4
A A
A 11-
A
‘ A A
40- A.
A A
30 T T T T I l
40 45 50 55 60 65 7O 75

Total Length (mm)

Figure 10: Maximum swimming velocities of 17 longnose dace.

56
than 60 mm Tl generally unable to burst faster than 50
cm/sec while dace greater than 60 mm T1 were able to burst

faster than 50 cm/sec.

DISCUSSION

Both adult and juvenile longnose dace were able to
utilize the fast channel in the habitat choice experiment.
The dace moved around in the fast channel by staying close
to the substrate surface and darting from rock to rock
(frequently pausing behind a rock before moving on). The 40
- 50 cm/sec currents in the fast channel were not
prohibitive to the juveniles of this size range, although,
by staying close to the substrate surface, the juvenile dace
may have been able to avoid prohibitive velocities. These
results confirm the observation made from the habitat use
survey (chapter one) that both the adults and juveniles
prefer the faster areas of the riffle (Figure 5). However,
the velocities in the fast channel were at the low end of
the range of velocities that make up the fast habitat
(higher velocities were not used because they inhibited
observations in the fast channel) and higher velocities
limit juvenile longnose dace in this size range (55 - 64 mm
Tl).

Even though the fry in this experiment were collected

57

from a nearby riffle, they preferred the slow channel over
the medium and fast channels. This supports the hypothesis
of Gibbons and Gee (1972) that longnose dace fry move out of
the quiet stream margins and into the riffles in response to
competition with blacknose dace fry in the stream margins.
The longnose dace fry may have also been inhibited by the
velocities in the faster channel, but, some fry did use the
fast channel (as evidenced by the error bar in figure 8).

The effect of adults on the shelter use of the
juveniles was to decrease the average time/use of the
shelter. In this experiment, the juveniles were more likely
to exit the shelter when the adults were present than when
they were absent. However, the results obtained here were
probably partly a function of the experimental design. Only
one shelter was available in the cages, whereas many
shelters are available in a normal riffle. The repeated
exit and re-entry observed by the juveniles when adults were
present would probably occur less frequently in a normal
riffle where the juvenile has a choice of shelters to use.
However, this study does demonstrate the significant point
that adult longnose dace can and will displace a juvenile
longnose dace from a shelter. This fact implies two costs
to juveniles coexisting with adults in the faster areas of a
riffle. First, every time a juvenile exits a shelter it
increases its risk of predation (there are few predators in

the faster areas of riffles, but, brook trout (Salvelinus

58

fontinalis) do forage in stream riffles and, historically,
were probably more abundant than at present). Second, every
time a juvenile exits a shelter it encounters the current.
Even if it is physically able to swim against the current,
it would still incur an energetic cost while looking for
another shelter.

If adult densities are high in a riffle, and, as has
been demonstrated, adults prefer the faster areas of a
riffle, a juvenile using the faster areas of the riffle runs
a high risk of being displaced from its shelter. There is
also a fair probability that the juvenile will encounter a
larger dace in the next shelter it finds, thereby increasing
the cost (both energetic and predation costs) to the
juvenile while it searches for an uninhabited shelter.
Thus, even though the resource (shelter) is not completely
limited in the faster areas of riffles, the probability of
being displaced from a shelter may be high enough, for
smaller longnose dace, that the gains of foraging in this
high food density area are more than offset by the energetic
costs of existing in the faster areas. In conclusion, even
though longnose dace develop a preference for the faster
areas of the riffle by the end of their first year of life,
they remain segregated from the adult longnose dace in those
faster areas. This segregation is probably maintained until
the juveniles grow to a size where the probability of being

displaced from a shelter in the faster areas by a larger

59
dace is low enough that the energetic costs of existing in
the faster areas no longer offset the gains of foraging in

these high food density areas.

CHAPTER THREE

RIFFLE MANIPULATIONS

60

61

INTRODUCTION

The results from the behavioral studies indicate that

the size related habitat segregation observed in the habitat

use survey may be the result of both intra-specific

competition for space in the faster areas of the riffle and

physical
longnose
dace are
riffle.
The

physical

limitations on hte swimming ability of smaller
dace. At smaller sizes (< 60 mm T1) the longnose

not physically able to use the faster areas of the

purpose of this study was to examine the roles that

constraints on the juveniles and intra-specific

competition with the adults play in determining the habitat

use of juvenile longnose dace. A series of whole riffle

manipulations of adult densities over a two year period was

conducted in order to assess the effects of adult densities

on the habitat use and diet of a single cohort of juveniles

as it grew through it's first two full summers.

METHODS

Study site:

This study was conducted during the summers of 1989 and

1990 on Twomile Creek, a third order stream in Michigan's

62
upper peninsula. The study site was located in Dickinson
County, T43N, R30W, about 500 m upstream of the confluence
of Twomile Creek and the Ford River. Three sequential
riffles were selected for the riffle manipulation study, and
a fourth was used for a study of the distribution of

invertebrates on rock surfaces within a riffle.

Riffle manipulations:

During the 1989 field season, adult and juvenile dace
densities were manipulated in three enclosed riffles in
order to examine the effects of adult density on the habitat
use of juvenile longnose dace. Three adjacent riffles in
Twomile creek were enclosed with % inch hardware cloth
fences (Figure 11), which were cleaned daily in order to
reduce their effect on the flow within the riffle. The
fences were placed in pools or runs between the riffles.
These fences were installed in mid-June after the late
spring floods had subsided. Fish traps (weir boxes) were
placed at the upstream and downstream ends of the study area
to capture all fish moving upstream and downstream through
the area. These traps were checked daily and all fish were
hand carried around the enclosed study area and released in
the direction of travel when captured.

Three treatments were used; 1) a control with no adults
or juveniles removed or added, 2) a low adult treatment with

all adults captured during a sample run removed, and 3) a

63

RIFFLE MANIPULATION DESIGN

 

 

 

 

Downstream 1’4 inch
Weir Box hardware cloth Upstream
Weir Box
pool / 26m pool
_ \ l . | VT"! ,
Y Control if Adult \5. Juvenile
10m [:1 riffle Removal i Removal I;
g riffle § riffle
<17
CURRENT

Figure 11: Design of riffle manipulations used in 1989 on
Twomile Creek

64

low juvenile treatment with all captured juveniles removed.
The riffles were 25 to 30 m long and each riffle varied in
width from 5 to 10 m. Each riffle was sampled with
electrofishing gear (using the procedure described in
chapter one) once a week for five weeks during early summer.
Data from the first sampling date (June 16) were on longnose
dace in the riffles before manipulation. After sampling on
that date, the juvenile and adult manipulations were
executed and maintained for the next four weeks. After
sampling on each date, the availability of three habitat
variables; depth, mean water column velocity and substrate
type was estimated using ten points on each of five evenly
spaced transects in each riffle (as described in chapter
one). On July 23, the fences were removed and normal fish
movements were allowed for 8 days. During this period, the
adult and juvenile longnose dace re-colonized the
experimental riffles (see table 6 for pre-manipulation
densities of adults and juveniles in each riffle). The
riffles were again enclosed on August 2, and the three
riffles were sampled before manipulation. Manipulations of
adult and juvenile densities were then conducted and
maintained for the next three weeks while the habitat use of
the longnose dace in each treatment was sampled weekly.

In 1990, this design was altered to eliminate the
potential confounding effects of inter-riffle differences.

Only one riffle (the juvenile removal riffle in 1989) was

65

used and manipulations of adult densities were conducted
sequentially within the riffle. The riffle was isolated in
the manner already described and manipulations were
conducted as follows: control (sample prior to manipulation,
the adults were removed immediatly after this sample) - low
adult (sample of juvenile habitat use in experimental riffle
one week after the adults were removed. Adults were planted
in the riffle immediatly after the low adult sample was
taken) - control (sample of adult and juvenile habitat use
taken one week after the adults were returned to the riffle,
the adults were removed immediatly after this sample was
taken) - low adult. Sampling occurred once each week for
four weeks in early summer and late summer. After sampling
the habitat use patterns of the adults and juveniles on a
control day, as many adults as possible were removed from
the riffle by making multiple passes through the riffle with
electrofishing gear. After sampling on a low adult density
day, adults (as many as were removed the previous week) were
captured from nearby riffles and placed within the enclosed
riffle in order to provide normal adult densities for the
following week. The first experiment lasted from June 20,
1990 to July 11. The second experiment lasted from August 1,
1990 to August 22. The enclosure was removed between
experiments.

Comparisons of the velocity, depth and substrate use

were made between all of the adults and juveniles captured

66

at the start of each manipulation experiment, and between
the juveniles captured in each treatment after manipulation.
Two dimensional contingency table analysis using chi-square
goodness of fit was used for tests of independence between
each of the habitat variables and size, and between each of
the habitat variables and treatments (for the juveniles).
Also, as described in chapter one, chi-square goodness of
fit was used to test for selectivity of each of the habitat

variables by the adults and juveniles in each treatment.

Diet analysis:

During the course of the 1989 riffle manipulation
studies, several longnose dace were collected for diet
comparisons between the juveniles and adults. At the start
of each experiment (June 21 and August 2), ten juveniles
were taken from both the control and adult removal riffles
and ten adults were taken from both the control and juvenile
removal riffles and preserved in 70% ETOH. These fish were
replaced with dace of the appropriate size class from the
river at large. At the end of each experiment, ten
juveniles and ten adult were taken from the control riffle
and ten juveniles were taken from the adult removal riffle
and ten adults were taken from the juvenile removal riffle.
All invertebrates in the guts were identified to family.
Invertebrates that occurred frequently (> 10%) in the guts

were divided into size groups based on head capsule lengths

67
(small = HCL < 0.5 mm., medium = 0.5 < HCL < 1.0 mm. and
large = HCL > 1.0 mm.). Schoener's (Schoener 1970) index of
dietary overlap was used to make comparisons between all
adults and juveniles collected prior to manipulation.
Comparisons were also made between juveniles from the
control and adult removal treatments.

During the 1990 riffle manipulation studies, twenty
adult and twenty juvenile longnose dace were collected on
the first day of sampling (control) for both experiments
(June 20 and August 1). The twenty juveniles were replaced
by twenty juveniles from a nearby riffle. Also, twenty
juveniles were collected for diet analysis on the first low
adult sampling day of each experiment. These were also
replaced by twenty juveniles from a nearby riffle.
Comparisons were made between adults and juveniles in the
high adult treatment, and between juveniles in the low adult

treatment and juveniles in the high adult treatment.

Invertebrate distribution on rock substrates:

The distribution of stream invertebrates on substrates
was examined at the end of each of the riffle manipulation
studies in 1989. On June 22 and on September 8, five
replicate samples of invertebrates on each of three velocity
exposures (substrate surfaces) were taken from rocks
(cobbles) in the fastest flowing areas of a riffle

(velocities ranged from 50 to 70 cm/sec). Substrate

68
surfaces were classified by their exposure to the current; F
= fully exposed (usually the front and/or top surface of the
rock), P = partially exposed (usually the sides of the rock)
and N = not exposed (usually the back of the rock). A
different rock was used for each exposure replicate (15
rocks were sampled on each date). Sampling of an exposure
surface consisted of selecting a rock surface of the desired
classification in the riffle (the riffle from which these
samples were taken is mostly composed of boulders and cobble
ranging in size from 15 to 30 cm maximum length and all
surfaces chosen were within that range of greatest
dimension), quickly picking the rock up out of the river and
holding the selected surface over a bucket while vigorously
brushing the invertebrates on that surface into the bucket.
Care was taken to prevent invertebrates from the other
surfaces of the rock from dropping into the bucket. The
invertebrates were preserved in 70% ETOH and later
identified to family or genus and common items were divided
into size classes using the system described above for diet
comparisons. For comparative purposes, the surface that was
sampled was covered with foil which was trimmed to match the
surface as closely as possible. This foil piece was then
marked to identify the sample with which it corresponded. A
leaf area meter was used to determine the surface area of
the foil, and therefore, the area of the surface sampled.

Comparisons of the invertebrate densities on each surface

69
were based on numbers/m2.

Kruskal-Wallis single factor analysis of variance by
ranks was used to compare the densities of each invertebrate
group on all three exposure surfaces (F, P and N). When
appropriate, a nonparametric Tukey-type multiple comparison
(Zar 1984) was used to determine which substrate surfaces

differed from each other.

RESULTS

Riffle Manipulations:

The juvenile removals conducted in the June and August
1989 riffle manipulations were unsuccessful. In June: 20,
26, 30, 31 and 38 juveniles were captured on successive
sampling dates, and in August 31, 35, 28 and 22 juveniles
were captured on successive sampling dates in the juvenile
removal riffle. The slight increase over time in June and
the slight decrease over time in August were also observed
in the other riffles indicating that the juveniles were able
to move through the hardware cloth fences unimpeded (more
adults than juveniles were captured in the wier boxes while
the studies were taking place, however due to gear
selectivity for larger fish, little can be said about
differential movement patterns between the two size groups).

As a result, the low juvenile treatment will be considered

70
to be a second control (control 2) riffle in this study.
Adult removal, on the other hand, significantly reduced
adult densities in the low adult treatment for three of the
four riffle manipulation studies (August 1989, and June and
August 1990) (Table 6). In the June 1989 study, the adult
manipulation was ineffective in the low adult treatment.
This riffle contained a low density of adults prior to
manipulation.

Comparisons of the three riffles used in the 1989
studies indicated that these three riffles were different
(Table 7). The control 2 riffle had more deep water, fast
water and boulders than the other two, while the other two
were well matched by depth and velocity but not by
substrate. In addition, the three riffles differed in their
habitat associations: in the control 1 riffle, substrate
size was associated with depth (boulders were found in
deeper water) and velocity (boulders were found in slower
water), in the control 2 riffle, substrate size was
associated with depth (boulders in deeper water) but
velocity was not associated with substrate size or depth, in
the low adult riffle, velocity and depth were positively
associated (faster velocities in deeper areas) but substrate
size was not associated with depth or velocity. Due to
these inter-riffle differences, the results of the 1989
riffle manipulations are difficult to interpret. The

protocol of the 1990 riffle manipulations was altered (as

71

Table 6: Juvenile and adult densities (numbers/m3) in each
treatment during each of the riffle manipulation studies in
1989 and 1990. The densities are corrected for estimated
capture efficiencies of 50 % for juveniles and 60 % for

adults (see chapter 1 for details on the determination of
capture efficiencies). The values in parentheses with the
1989 data indicate the pre-manipulation densities for each

treatment.

studies are the control densities.

Pre-manipulation densities for the 1990

 

 

 

 

 

Date Density (#/m2)

Treatment Juveniles Adults
June 1989

Control 1 0.652 (0.420) 0.180 (0.203)
Control 2 0.376 (0.238) 0.265 (0.085)
Low adult 0.706 (0.268) 0.118 (0.100)
August 1989

Control 1 0.490 (0.498) 0.125 (0.168)
Control 2 0.406 (0.438) 0.225 (0.353)
Low adult 0.456 (0.448) 0.068 (0.192)
June 1990

Control 0.182 0.188

Low adult 0.214 0.053

August 1990

Control 0.268 0.187

Low adult 0.182 0.118

 

72

Table 7: Comparison of the habitat availabilities of each
habitat variable in each of the three riffles used in the
1989 riffle manipulation studies. The data are the
proportions of each habitat variable in each riffle.

 

 

 

Riffle
Habitat variable Control 1 Low adult Control 2
Shallow (< 15 cm) 0.51 0.48 0.42
Deep (> 15 cm) 0.49 0.52 0.58
Slow (< 40 cm/sec) 0.86 0.88 0.69
Fast (> 40 cm/sec) 0.14 0.12 0.31
Boulder 0.19 0.30 0.40

Other 0.81 0.70 0.60

 

73
discussed in methods) to eliminate inter-riffle effects.

In the June 1989 manipulation, the adults and juveniles
were not segregated by depth in the three riffles before
manipulation or in the control 1 riffle after manipulation
(Table 8). However, the adults and juveniles were
segregated by depth in the control 2 riffle, and in the low
adult riffle. In the control 2 riffle, the juveniles
selected shallow water (Figure 12), while in the control 1
riffle the juveniles were not selective of depth, and before
manipulation, the juveniles selected mid range depths In the
low adult riffle, the juveniles did not select a particular
depth range but did avoid deep areas. There is no apparent
pattern to suggest an adult affect on the depth selection of
the juveniles. Contingency table analysis of the depth use
of the juveniles in the low adult riffle versus the depth
use of the juveniles in the control 1 riffle (Table 9) and
versus the depth use of the juveniles in the control 2
riffle (Table 10) indicated that the depth use was similar
across treatments and was not affected by adult densities.
The different selection patterns illustrated in Figure 11
may reflect inter-riffle differences in availability. For
instance, about 42% of the control 2 riffle was shallow,
compared to 51% for the control 1 riffle and 48% for the low
adult riffle. Since the juveniles in each treatment used
the shallow areas in similar proportions (Tables 9 and 10),

they showed stronger selection for shallow areas in the

74

Table 8: Contingency table analysis (using chi-square goodness of
fit) of first order interactions between size and each of the
habitat variables in unmanipulated (Initial) and manipulated

riffles in the June 1989 riffle manipulation study.

The sign

next to chi-square values greater than 0.5 indicates the

direction of the difference (observed - expected).

 

 

Size Depth Velocity Substrate Type
Shallow Deep Slow Fast Boulder Other
Before
Juveniles
Observed 16 186 160 42 58 144
chi-square 0.04 0.00 l.69+ 3.33— 1.36— 0.69+
Adults
Observed 8 78 48 37 38 47
Chi-square 0.09 0.01 3.00- 7.90+ 3.25+ 1.63-
Significance NS p < 0.001 p < 0.025
Control 1
Juveniles
Observed 238 304 430 112 96 446
chi-square 0.00 0.00 0.29 0.82- 5.62- l.64+
Adults
Observed 65 83 105 45 60 90
chi-square 0.00 0.00 1.04- 3.56+ 20.31+ 5.91-
Siqnificance NS p < 0.025 p < 0.001
Low Adult
Juveniles
Observed 252 330 474 108 216 366
chi-square 2.33+ 1.51- 0.94+ 3.30- 0.96- 0.63+
Adults
Observed 15 82 55 42 53 43
chi-square 14.01- 9.06+ 5.61- 19.83+ 5.83+ 3.83-
Siqnificance p < 0.001 p < 0.001 p < 0.001
Control 2
Juveniles
Observed 118 132 202 48 110 140
chi-square 5.46+ 3.36- 3.29+ 8.11- 3.05- 3.30—
Adults
Observed 48 138 108 78 117 70
Chi-square 7.34- 4.51+ 4.43- 10.89+ 4.08+ 4.41-
Significance p < 0.001 p < 0.001 p < 0.001

 

75

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20
\ ’ \
before -
15~ Juveniles P < 0001
g m, E]
'6 Adults P < 005
>
o 5-
a
3 1 1
‘4’
8 -6-
.10.
‘5 0-9 ‘ 10-19 ' 20-29 Y >33
‘
control 1 Juveniles NS
0 5- C3
.3 Adults P < 0005
>
I)
a 0- m—
D
U
m
E
0 -5-
~10 0.9 T 10-19 . 20'29 V ’m
10
' \‘.
control 2
Juvenile: P < 001
s« [:l

Adults P < 0&5

 

 

 

 

Chi-Square Value
C?

 

 

 

 

 

 

 

 

 

 

 

“0 0.9 ' 10-19 V 20-29 . >30
low adult m
0

D
2
g .9 SS
g Juveniles P < 0001
3
q) 40—4
E
0

.15-4

'29 0-9 ' 10.19 ' 20-29 T >30

Depth Cell (cm)

Figure 12: Chi-square analysis of the depth distribution of
juvenile and adult longnose dace captured in each
treatment during the June 1989 riffle
manipulation study.

76

Table 9: Contingency table analysis (using chi—square goodness
of fit) of first order interactions between the control 1 and
low adult treatments and the use of each of the habitat variables
by the juveniles in each treatment for each riffle manipulation
study in 1989. The sign next to chi-square values greater than
0.5 indicates the direction of the difference (observed -
expected).

 

 

Size Depth Velocity Substrate type
Shallow Deep Slow Fast Boulder Other
June 1989
Low Adult
Observed 252 330 474 108 216 366
chi-square- 0.01 0.00 0.07 0.31 l8.3l+ 7.04-
Control 1
Observed 238 304 430 112 96 446
chi-square 0.01 0.00 0.08 0.33 19.57- 7.56+
Significance NS NS p < 0.001

August 1989

Low Adult
Observed 126 144 224 46 88 182
chi-square 4.29- 5.49+ 3.03+ 8.57- 12.84+ 3.68-
Control 1
Observed 196 108 200 104 40 264
chi-square 3.81+ 4.87- 2.69- 7.62+ 11.39— 3.27+

Significance p < 0.001 p < 0.001 p < 0.001

 

77

Table 10: Contingency table analysis (using chi-square
goodness of fit) of first order interactions between treatments
and the use of each of the habitat variables by the juveniles in
each treatment for each riffle manipulation study in 1989 and
1990. The sign next to chi-square values greater than 0.5
indicates the direction of the difference (observed — expected).

 

 

Size Depth Velocity Substrate type
Shallow Deep Slow Fast Boulder Other
June 1989
Low Adult
Observed 252 330 474 108 216 366
chi-square 0.17 0.14 0.00 0.00 0.64— 0.40
Control 2
Observed 118 132 202 48 110 140
chi—square 0.42 0.33 0.00 0.04 l.46+ 0.94-
Significance NS NS NS
August 1989
Low Adult
Observed 126 144 224 46 88 182
Chi-square 0.12 0.12 0.30 1.18- 0.08 0.06
Control 2
Observed 86 84 128 42 60 110
chi-square 0.20 0.18 0.48 l.88+ 0.14 0.06
Significance NS p = 0.05 NS
June 1990
Low Adult
Observed 30 24 28 26 22 32
chi-square 0.56 0.52 2.36- 5.32+ 0.10 0.06
Control
Observed 36 46 66 16 30 52
Chi-square 0.36 0.34 1.56+ 2.36- 0.06 0.04
Significance NS p < 0.001 NS
August 1990
Low Adult
Observed 14 58 30 42 36 36
Chi-square 0.16 0.04 2.16- 2.58+ 0.08 0.06
Control
Observed 26 86 70 42 52 60
Chi-square 0.10 0.02 1.40+ 1.66- 0.04 0.04
Significance NS p < 0.01 NS

 

78
control 2 riffle where the shallow areas were less
abundant.

In the August 1989 riffle manipulation study, the
adults and juveniles were not segregated by depth in the
three riffles before manipulation but they were segregated
by depth in each of the riffles after manipulation (Table
11). In all cases except the control 2 riffle, the
juveniles selected mid range depths, in the control 2
riffle, the juveniles did not select any specific depth
range (Figure 13). A comparison of the depth use of the
juveniles in the low adult riffle versus the depth use of
the juveniles in the control 1 riffle indicated that the
distributions were different (Table 9), however, there was
no difference in the depth use of the juveniles in the low
adult treatment and the juveniles in the control 2 treatment
(Table 10). The juveniles in the control 1 riffle used the
shallow areas in a higher proportion than the juveniles in
the low adult riffle (Table 9), but the selection patterns
for these two riffles were similar (Figure 13). This also
appears to be due to inter-riffle differences in
availability. About 41% of the control 1 riffle was between
10 and 19 cm deep, while only 27% of the low adult riffle
fell within that range. Even though a lower proportion of
the juveniles in the low adult riffle used this depth range
(51%) than did the juveniles in the control 1 riffle (60%),

the lower proportion of the available habitat that fell

79

Table 11: Contingency table analysis (using chi-square
goodness of fit) of first order interactions between size and
each of the habitat variables in unmanipulated (Initial) and
manipulated riffles in the August 1989 riffle manipulation study.
The sign next to chi-square values greater than 0.5 indicates the
direction of the difference (observed - expected).

 

 

Size Depth Velocity Substrate type
Shallow Deep Slow Fast Boulder other
Before
Juveniles
Observed 118 150 210 58 96 172
chi-square 0.60+ 0.42 0.68+ 1.93- 0.44 0.27
Adults
Observed 42 80 78 43 53 68
chi-square 1.31- 0.9l+ 1.50- 4.29+ 0.97+ 0.60-
Significance NS p < 0.005 NS
Control 1
Juveniles
Observed 196 108 200 104 40 264
Chi-square 0.94+ 1.42- 0.00 0.00 0.60- 0.10
Adults
Observed 35 45 53 27 17 62
chi-square 3.57- 5.38+ 0.00 0.00 2.29+ 0.40
Significance p < 0.001 NS NS
Low Adult
Juveniles
Observed 126 144 224 46 88 182
chi—square 0.90+ 0.67- 0.37 1.44- 0.20 0.10
Adults
Observed 7 33 23 17 18 22
chi-square 6.05- 4.56+ 2.48- 9.78+ 1.35+ 0.70-
Significance p < 0.001 p < 0.001 NS
Control 2
Juveniles
Observed 86 84 128 42 60 110
chi-square 5.99+ 3.81- 4.66+ 7.68- 3.26- 2.61-
Adults
Observed 17 78 37 58 58 37
chi-square 10.73- 6.82+ 8.33- 13.774 5.83+ 4.68-
Significance p < 0.001 p < 0.001 p < 0.001

 

80

 

 

 

 

 

 

 

 

 

 

 

2"
,5. before mules P < 0031
o I:
S 10‘ mp<0001
>
3 s—
S
‘t’
E
0
4 0-9 ' 10-19 I 20-29 ' >30
ac
ED
‘6‘ controll JuvenilesP<0oo1

Chi-Square Value

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.9 ' 10-19 ' 20-29 >31
1
& contolz
D
3
. w:
>
0
g 6-
(‘3 JuvenilesNS
- .10—
6 E3
AdultsP<0m1
.15—
0-9 ‘ 10-19 ' 20-29 ' >30
low adult , \

Chi-Square Value

JIMDIIOS P < QCDI

 

 

 

 

0.9 ' 10-19 ' 20.29
DeptnCell(cm)

Figure 13: Chi-square analysis of the depth
juvenile and adult longnose dace
treatment during the August 1989
manipulation study.

 

>3)

distribution of
captured in each
riffle

81
within that range combined with the lowered proportion of
juvenile use of that range to produce a selection pattern
that was similar to the pattern exhibited in the control 1
riffle. The difference in the depth use of the juveniles in
the low adult riffle and the control 1 riffle may be related
to adult densities, however, there was no difference in the
depth use of the juveniles in the low adult riffle and the
control 2 riffle (Table 10), indicating that adult densities
had no affect on the depth use of the juveniles.

In the 1990 riffle manipulations, the adults and
juveniles were segregated by depth in the control treatments
in June but not in August (Table 12). Two days prior to the
last sampling day of the June manipulation (a low adult
sample), a storm event raised the water levels and washed
out a portion of the weirs allowing the adults into the
experimental riffle. The data from this date were not used
due to the high adult densities and the high water
velocities. In June, the juveniles were found in shallow
waters in a higher proportion than the adults (Table 12).

In the control treatment, the juveniles were not selective
of depth (Figure 14), but they selected shallower areas (<
10 cm deep) in the low adult treatment. However, this
selection was weak and contingency table analysis of the
depth use of the juveniles in the low adult treatment versus
the depth use of the juveniles in the control treatment

indicated that there was no difference in the depth use of

Table 12:

82

Contingency table analysis (using chi-square

goodness of fit) of first order interactions between size and
each of the habitat variables in unmanipulated riffles at the
start of each riffle manipulation study in 1990.
to chi-square values greater than 0.5 indicates the direction of
the difference (observed - expected).

The sign next

 

 

Size Depth Velocity Substrate type-
Shallow Deep Slow Fast Boulder Other
June 1990
Juveniles
Observed 36 46 66 16 30 52
chi-square 6.00+ 2.48- 7.78+ 10.39— 1.32- l.09+
Adults
Observed 15 77 33 58 48 43
chi-square 5.33- 2.22+ 7.00- 9.38+ 1.20+ 0.98-
Significance p < 0.001 p < 0.001 p < 0.05
August 1990
Juveniles
Observed 26 86 70 42 52 60
Chi-square 0.34 0.09 1.53+ 1.79- 3.60- 5.48+
Adults
Observed 13 63 32 45 62 15
Chi~square 0.50- 0.13 2.22- 2.60+ 5.24+ 7.95-
Significance NS p < 0.005 p < 0.001

 

83

 

control Juveniles NS

I:

Adults P < 0.025

 

10"

 

 

&

 

 

 

Chi-Square Value

 

 

 

 

0-9 fl 10-19 1 20-29 >30

 

\‘
low adult \\\\
Juveniles P < 005

5—4

 

 

Chi-Square Value

 

 

 

M

0-9 Y 10-19 1 20-29 ' > :30
Depth Cell (cm)

Figure 14: Chi-square analysis of the depth distribution of
juvenile and adult longnose dace captured in each
treatment during the June 1990 riffle
manipulation study.

84
the two groups of juveniles (Table 10). In August of 1990,
the juveniles in both treatments avoided shallow areas (< 10
cm deep) although not significantly so in the control
treatment (Figure 15). Contingency table analysis of the
depth use of the juveniles in each treatment indicated that
adult removal had no effect on the depth use of the
juveniles (Table 10).

The juveniles and adults were segregated by substrate
size in all treatments in the June 1989 riffle manipulation
(Table 8). In all cases, the adults selected boulders
(Figure 16) while the juveniles were either non-selective of
substrate size (before manipulation and control 2), avoided
large substrates (control 1) or selected large substrates
(low adult). The substrate use of the juveniles in the low
adult riffle was not significantly different than that of
the juveniles in the control 2 riffle (Table 10) but was
different than the substrate use of the juveniles in the
control 1 riffle (Table 9). In the August 1989 riffle
manipulation, the juveniles and adults were segregated by
substrate type in the control 2 riffle (Table 11) but not in
the other riffles. In all cases, the juveniles were not
selective of either substrate type (Figure 17), although
there was a trend toward avoidance of boulders by the
juveniles in the control 1 riffle. The substrate use of the
juveniles in the low adult riffle was similar to that of the

juveniles in the control 2 riffle (Table 10) but different

85

 

1G

\\\\\‘
control Juveniles NS

5‘ I:

Adults P < 0.025

 

 

 

Chi-Square Value
‘?

 

 

o-e ' 10-19 ' 20-29 ' >30

Juveniles P < 0.025

 

low adult

 

 

Chi-Square Value

 

 

 

0-9 ' 10-19 ' 20-29 ‘ > 30
Depth Cell (cm)

Figure 15: Chi-square analysis of the depth distribution of
juvenile and adult longnose dace captured in each
treatment during the August 1990 riffle
manipulation study.

86

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

it}
before ;\
wwwuNS
O
S e [I]
2 Adults P < 0.01
a
8
9
:' BOULDER ' OTHER
'01
control 1 \
‘0‘ Juvenile: P < 0m1 *3
g E]
3 Adults P < 0.05
> 5~
5; m
a.) 0"
5 I__ i
-5-1
40 .
BOULDER OTHER
‘5‘ control 2
Juveniles NS
0
s "r i:
> 5‘ Adults P < 0001
O
a
3
0'
9
5 5
40-
15 .
BOULDER omen
15
low adult \\\\‘
10" Juveniles P < 001
E?
>
O
a
a
g SSSSSS
O
-5-1
-1 Y
BOULDER OTHER

Substrate Type

Figure 16 Chi-square analysis of substrate use patterns of
juvenile and adult longnose dace captured in each
treatment during the June 1989 riffle
manipulation study.

87

 

Ch

Chi-Square Value
9

 

before
Juveniles NS

E3

AOUIISP<0C5

““1—

 

 

 

 

 

 

In

BOULDER OTHER

 

(h

Chi-Square Value
9

uxmd1 SN
Juveniles NS

E]

Adults NS
W

 

4n

BOULDER OTHER

 

10

‘l‘

Chi-Square Value

control 2

Juveniles NS

C]

Adults P < 01131

 

 

 

 

 

 

 

 

 

 

 

Chi-Square Value
C)

 

 

 

" BOULDER OTHER
low adult \\\\
Juveniles NS
" BOULDER T OTHER

Substrate Type

Figure 17: Chi-square analysis of substrate use patterns of

juvenile and adult longnose dace captured in each
treatment during the August 1989 riffle
manipulation study.

88
to that of the juveniles in the control 1 riffle (Table 9).
In both the June and August riffle manipulation studies, the
juveniles in the control 1 riffle displayed some degree of
avoidance of boulders (Figures 16 and 17). This may be
related to inter-riffle differences in habitat variable
associations. In the control 1 riffle (and not in the other
two riffles), boulders were associated with slow velocities
and (as discussed below) the juveniles in this riffle used
faster velocities than did the juveniles in the control 2
and low adult riffles.

The adults and juveniles were segregated by substrate
type in the control treatments of both the June and August
1990 studies (Table 12). In both cases, the adults selected
boulders and the juveniles were not selective of substrate
type (Figures 18 and 19). Contingency table results from
the analysis of the substrate use of the juveniles in the
low adult treatments versus that of the juveniles in the
control treatments for both studies were non-significant,
indicating that the adult removal had no effect on the
substrate use of the juveniles (Table 10).

During the June 1989 riffle manipulation study, the
adult and juvenile longnose dace were segregated by velocity
before manipulation and in all treatments after manipulation
(Table 8). Before manipulation and in the control 1 riffle,
the adults did not select a particular velocity range, but

they selected faster velocities in the control 2 riffle

89

 

1G

 

control

 

Juveniles NS

l:l

Adults P < 0.001

 

 

Chi-Square Value

 

 

 

 

 

01 ill

BOULDER OTHER

 

low adult

Juveniles NS

 

Chi—Square Value

 

 

(ill

Figure 18: Chi-square analysis of substrate use patterns of
juvenile and adult longnose dace captured in each
treatment during the June 1990 riffle

BOULDER ' OTHER
Substrate Type

manipulation study.

 

 

90

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

35
3°“ control \\\\\\\j
5‘ Juveniles P < 0.025
% 2" 1:1
> 16- Adults P < 0.001
S 10-
5
° TN
9 m
E
o -&
-10—
.15—
-m '
BOULDER OTHER
low adult \\\\‘
o Juveniles NS
2
C1!
> m
d.)
a
3 m
U
9
E
o
.5 7
BOULDER OTHER
Substrate Type

Figure 19: Chi-square analysis of substrate use patterns of
juvenile and adult longnose dace captured ln each
treatment during the August 1990 riffle
manipulation study.

91
(Figure 20). In all cases, the juveniles selected mid-range
velocities and avoided high velocities. Contingency table
analysis of the velocity use of the juveniles in the low
adult riffle versus the velocity use of the juveniles in the
control 1 riffle (Table 9) and versus the velocity use of
the juveniles in the control 2 riffle (Table 10) indicated
that adult densities had no effect on the velocity use of
the juveniles. In August of 1989, the adults and juveniles
were segregated by velocity before manipulation and in the
low adult and control 2 riffles after manipulation (Table
11). In the control 1 riffle, however, there was no size
related segregation by velocity. Contingency table analysis
of the velocity use of the juveniles in the low adult riffle
versus the velocity use of the juveniles in the control 1
riffle (Table 9) and the velocity use of the juveniles in
the control 2 riffle (Table 10) indicated that the juveniles
in the low adult riffle used velocity differently than the
juveniles in the control riffles. In both cases, the
juveniles in the low adult riffle used slower velocities
than the juveniles in the control riffles. The juveniles in
the low adult riffle selected velocities between 30 and 49
cm/sec, while in the control 1 riffle they strongly selected
velocities between 40 and 49 cm/sec and in the control 2
riffle they were not selective of velocity (Figure 21).
However, velocities > 40 cm/sec in the control 2 riffle were

much more abundant (31%) than they were in the low adult

92

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

as
2‘” before 5 \\\\
Juveniles P < 0001
3 ma
9
. 10~
. .l
3
0'
‘I’
E ‘L
o u
-6~
1 e . - 4
0 9 10-19 20 29 30-39 40 49 >50
,& coNMNi NS
104 Juveniles P < 0001
3 C3
; 5‘ AdultsNS
O
a G .-
3
O’
‘9 -5-1
E §7
0 -10~
-15-
0-9 1 10-19 ' 20.29 ' 30.39 ' 40-49 ‘ >50
15
cmumuz §§
10.. Juveniles P < 001
D
g C]
2 5. AOUlt5P<0EnS
. N
3
(Cpl' G S‘J u j
E l I
o
-5—
"G V v v v ‘7’
0.9 10-19 2029 11.39 40-49 >50
20
15‘ lowadult \\\\\~
Juveniles P < 0001
3 10~ .
a: \‘
> \
2 9 §§
GO
3 \\
to V
E .5-4 § \\\\
o
-10~
15

 

0-9 ' 10-19 ' 20-29 ' 30-39 ' 40-49 ' >50
Velocity Cell (cm/sec)

Figure 20: Chi-square analysis of velocity distributions of
juvenile and adult longnose dace captured in each
treatment during the June 1989 riffle
manipulation study.

Chi-Square Value

93

 

before
Juveniles P < 0(25

10" D

AdultsP<Qm5

Chi-Square Value
‘1‘

 

 

 

§§_J IsiTiugjflk—xigls:: md:]

 

 

 

 

 

 

" o-e ' 10-19 ' 20.29 ' ao-as_ ‘ 40-49 ' >50
11‘
mxmdl

§§
75‘ JuvenilesP<0001
s- C3

Adultp<uoo1
x-l
15«
‘5‘ N —‘— —
'5 v v v v W

0-9 10-19 zo-m 30-39 ‘049 >50

10

cmflnflz

‘1‘

 

 

 

Juveniles NS

C3

Adults P < 0001

‘Cnl-Square Value
C)

in
1

v

0.9 ' 10.19 ' 20-29 ' ao-ae r 40-49 >50

 

 

low adult \\
Juveniles P < 001

Wm§ §s _

10-19 ' 20-29 ' 30-39 * 40-49 ' >50
Velocity Cell (cm/sec)

 

Chi-Square Value
‘f‘

 

 

 

(h

Chi-square analysis of velocity distributions of
juvenile and adult longnose dace captured in each
treatment during the August 1989 riffle
manipulation study.

Figure 21:

94
riffle (12%), accounting for the different selection
patterns seen between these riffles. The juveniles in the
control 2 riffle used these faster velocities in a higher
proportion than the juveniles in the low adult riffle (Table
10), but since the faster velocities were more scarce in the
low adult riffle, the juveniles in the low adult riffle
showed stronger selection for faster velocities (Figure 21).
Since the juveniles used slower velocities in the low adult
riffle than in the other riffles, and adults generally use
higher velocities than the juveniles, the difference in the
juvenile velocity use observed between the low adult riffle
and the two control riffles was not related to adult
removal.

The adults and juveniles were segregated by velocity in
control treatments during both of the 1990 riffle
manipulations (Table 12). Contingency table analysis of the
velocity use of the juveniles in the low adult treatments
versus that of the juveniles in the control treatments for
both studies indicated that the juveniles responded to adult
removal by moving into faster water (Table 10). In June
1990, this move involved a change in selection from 20 - 39
cm/sec with adults present to non-selective in the absence
of adults (but with a trend towards selecting velocities
between 40 and 49 cm/sec) (Figure 22). In August 1990, for
the first time in this study, the juveniles responded to

adult removal by moving into the fastest areas of the riffle

95

 

 

 

 

Chi-Square Value

 

 

 

 

 

 

ac
control t\\
‘5‘ Juvenlles P < 0.001
,0: I:
Adults P < 0.025
5-
" “I l LJ
.6—
-1c 1 . f . .
0-9 10-19 20-29 ao-ao 40-49 >50
low adult ‘
o Juveniles NS
2
CU
> \
0 t
a __ ._ _.
3
O’
m. §7
E
o

 

 

 

ch

0-9 ' 10-19 ' 213-29 ' 30-39 ' 40-49 ' >50
Velocity Cell (cm/sec)

Figure 22: Chi-square analysis of velocity distributions of
juvenile and adult longnose dace captured in each
treatment during the June 1990 riffle
manipulation study.

96
(Figure 23).

The design of the riffle manipulation studies in 1989
was plagued by inter-riffle differences in habitat
availability (Table 7) and by ineffective manipulations of
adult densities in the June study (Table 6). As a result,
few conclusions can be made from this data. However, by
comparing the habitat use of the juveniles in the low adult
riffle after manipulation with the habitat use of the
juveniles in that riffle before manipulation for the August
1989 study (adult density manipulations were ineffective for
the June study but reduced the adult densities by about 65%
in the August study), it is possible to assess the effects
of the adults on the habitat use of the juveniles for that
period. There was no change in the depth use and the
substrate use of the juveniles in this riffle after
manipulation (Table 13), however, there was a change in the
velocity use. The juveniles used faster velocities before
the adults were removed than they did after the adults were
removed. Since velocity segregation between the size
classes involved the adults using higher velocities than the
juveniles, adult removal does not explain this shift into
slower water by the juveniles after the adults were removed.
Adult longnose dace had no apparent effect on the habitat
use of the juveniles in the August 1989 riffle manipulation

study.

97

 

   

Chl-Square Value

 

 

 

 

 

 

 

 

Chi-Square Value

 

 

 

 

 

 

 

20
control .
159 \: Juveniles P < 0.001
m
mg l:
10— § Adults P < 0.025
g
s- § fl
0 __, g\ _1—J
6 0-9. ' 10-19 V 210-29 30 an 40-49 >50
m .
low adult
15* Juveniles P < 0.05
104
5“ §
0 g
g
J5 0-9 ' 10-19 ' 20-29 Y 30-39 ' 40-49 Y >50

Velocity Cell (cm/sec)

Figure 23: Chi-square analysis of velocity distributions of
juvenile and adult longnose dace captured in each
treatment during the August 1990 riffle
manipulation study.

98

Table 13: Contingency table analysis (using chi-square
goodness of fit) of first order interactions between juvenile
habitat use in the low adult riffle before to manipulation of
adult densities in August 1989 and juvenile habitat use in that
riffle after adult removal. The sign next to chi-square values
greater than 0.5 indicates the direction of the difference
(observed - expected).

 

 

Size Depth Velocity Substrate type
Shallow Deep Slow Fast Boulder Other
Before
Observed 48 50 . 72 26 38 6O
chi-square 0.06 0.06 0.59- 2.41+ 0.58+ 0.30
After 1
Observed 126 144 224 46 88 182
chi-square 0.02 0.02 0.21 0.88- 0.21 0.11

Significance NS p < 0.05 NS

 

99
Diet analysis:

The dietary overlaps between the juveniles and adults
captured at the start of each riffle manipulation study were
relatively low (Table 14). In some cases, the adult diets
were broader, but in general the differences arose from
varying proportions of the major diet items of juveniles and
adults. For instance, in all cases, Chironomidae made up
the major component of the juveniles diet (from 57 to 74%),
while Chironomidae made up a much smaller proportion of the
adults diet (from 6 to 39%). In addition, although
Simulidae occurred in the diets of both the adults and the
juveniles on two dates, Simulidae constituted a larger
portion of the adults diet than of the juveniles diet (Table
14).

Under the original study design in 1989, the juveniles
for between treatment diet comparisons of juveniles in the
low adult treatment versus juveniles in the control
treatment were taken from the control 1 riffle. As a
result, there is no data available for comparing the diets
of the juveniles in the low adult treatment with the diets
of the juveniles in the control 2 treatment. The guts of
the juveniles collected after manipulation in the June 1989
study were nearly all empty. Therefore, inter-treatment
comparisons of the diets of the juveniles cannot be made for
the June 1989 study. The diets of the juveniles in the low

adult treatment in the August 1989 study were fairly

100

Table 14: Dietary overlaps (Schoener 1970) between adult and
juvenile longnose dace captured in unmanipulated riffles at the
start of each riffle manipulation study in 1989 and 1990. B =
Baetis, C = Chironomidae, G = Glossosoma, H = Hydropsychidae,

Hp = Heptageniidae, L = Leucotrichia, P = Psychomyia, R =
Rhyacophilidae and S = Simuliidae. Subscripts refer to size
class based on head capsule lengths where; S = O - 0.5 mm, M =
0.5 - 1.0 mm and L = > 1.0 mm. Values in parantheses indicate the
percentage of the food items in the diets of the juvenile and
adult dace.

 

Major Diet Items

 

 

Date NJ lg Overlap Juveniles Adults
June 1989 22 19 55% C (61) C (39)
P (11) P (7)
B (17) H“L (5.15)
HP (13)
St" (6.6)
August 1989 20 17 48.5% C (64) C (19)
85(8) SS. (37,9)
L (10) G '28)
June 1990 11 10 30% C (74) C (6)
ss'M (9,12) 85” (51,38)
August 1990 17 18 40% C (57) C (11)
B (11) B (15)
p (17) R (13)
sSH (22,7)

Hsj, (6.6)

 

101
different (overlap = 61.5%) (Table 15) from the diets of the
juveniles in the control 1 riffle. The juveniles in the
control 1 riffle had a broader diet than the juveniles in
the low adult riffle, including more Simuliidae. This
difference is probably due to the fact that the juveniles in
the control 1 riffle used faster areas than the juveniles in
the low adult riffle (Table 9). The diets of the juveniles
in the control treatment in June 1990 were similar to the
diets of the juveniles in the low adult treatment (Table
15), while the diets of the juveniles in the control
treatment in August 1990 were substantially different than
the diets of the juveniles in the low adult treatment. The
main difference in the August diets was the inclusion of
Simulidae into the diets of the juveniles in the low adult

treatment (Table 12).

Invertebrate distribution on rock substrates:

Total invertebrate densities on the three exposure
surfaces were not different on either of the sampling dates.
There were, however, differences for some of the
invertebrate groups (Table 16). Simulidae and
Hydropsychidae were more abundant on the exposed surfaces in
the early summer and only small Hydropsychidae were more

abundant on exposed surfaces in late summer.

102

Table 15: Dietary overlaps (Schoener 1970) between juvenile
longnose dace captured in low adult density and control
treatments during the riffle manipulation studies in 1989 and
1990. B = Baetis, C = Chironomidae, H = Hydropsychidae, L =
Leucotrichia, P = Psychomyia and S = Simuliidae. Subscripts
refer to size classes based on head capsule lengths where S = O -
0.5 mm and M = 0.5 - 1.0 mm. Values in parentheses indicate the
percentage of the food items in the diets of the juvenile dace in
each treatment.

 

Major Diet Items

 

 

Date NLA NC Overlap Low Adult Control
June 1989 too many empty guts to permit comparison
August 1989 10 9 62% C (69) C (38)

L (9) SEN (14, 6)

HS," (6.6)

June 1990 12 ll 72% C (50) C (74)

S& (9,8) SSA (9,12)

3 (28)
August 1990 12 17 56% C (23) C (57)

B (17) B (11)

P (6) P (17)

85(24,6)

 

103

Table 16: Results of Kruskall-Wallis comparisons of
invertebrate densities on each of three exposure surfaces (N
= not exposed, P = partially exposed and F = fully exposed).

 

SIGNIFICANT COMPARISONS

 

N vs P N vs F P vs F
7/22/89
small Hydropsychidae medium Hydropsychidae none

total Hydropsychidae large Hydropsyhchidae
total Simuliidae

9/8/89

small Hydropsychidae none none

 

' all significance levels are P < 0.05 and in all cases N
was less than either P or F.

104

DISCUSSION

Due to the inter-riffle differences in the 1989 riffle
manipulations, only a few conclusions can be made from these
studies. Specifically: 1) There was no evidence to suggest
an adult effect on the depth use of the juveniles. Juvenile
depth use was consistent across treatments in June and
varied between treatments in August, but did not change in
the low adult riffle after the adults were removed in August
(Table 13). 2) There was no evidence to suggest an adult
effect on the substrate use of the juvenile longnose dace.
There was a difference in the boulder use of the juveniles
in the low adult riffle and the boulder use of the juveniles
in the control 1 riffle, but this appeared to be related to
inter-riffle differences in boulder availability and habitat
associations. The boulder use of the juveniles in the low
adult riffle did not change after the adults were removed
(Table 13). 3) There was no evidence to suggest an adult
effect on the velocity use of the juveniles. The juveniles
in the low adult riffle reduced their velocity use after the
adults were removed (Table 13), but, since the adults use
higher velocities than the juveniles, this does not appear
to be a response to adult removal. 4) Regardless of adult
densities, the juveniles avoided faster velocities (> 50

cm/sec) in June and did not select faster velocities in

105
August (Figures 20 and 21).

The study design used in 1990 eliminated the problem of
inter-riffle differences and demonstrated that the adults
had no effect on the depth and substrate use of the
juveniles, but did affect the velocity use of the juveniles
(Figures 22 and 23, Table 10).

There was no consistent pattern of depth segregation
exhibited in this study, but there was a fairly consistent
pattern of substrate segregation. With the exception of the
August 1989 study, the adults were always associated with
larger substrates than the juveniles (Tables 1, 8, and 12).
In the August 1989 study (Table 11), segregation did not
occur before manipulation or in two of the riffles after
manipulation. However, the pattern of the + and -'s in the
contingency tables for these riffles indicated that there
was a tendency for the adults to use larger substrates.
Adult removal had no effect on the substrate use of the
juveniles indicating that substrate segregation resulted
from some factor(s) other than intra-specific competition.
Larger substrates may harbor predators (brook trout,

Salvelinus fontinalis and/or burbot, Lota lota) of small

 

longnose dace. However, in most cases the juveniles
displayed some degree of selection for larger substrates,
indicating that predation risk on the juveniles probably
does not explain segregation. Alternatively, larger

substrates may provide shelter for larger longnose dace from

106
avian or terrestrial predators and longnose dace may develop
a preference for larger substrates as they grow.

There was a consistent pattern of velocity segregation
between the adults and juveniles throughout the course of
this study. The effect of adult longnose dace on the
velocity use of juvenile longnose dace appeared to be a
function of juvenile size. Throughout the period of this
study, the adults and juveniles in control or un-manipulated
riffles were segregated by mean water column velocity (with
the exception of the control 1 riffle in August 1989)
(Tables 8, 11 and 12). Low adult densities in June 1989
(mean size of the juveniles was 51.7 mm Tl) had no effect on
the velocity use of the juveniles (Table 8, Figure 20). In
August 1989, the juveniles (mean size = 56.7 mm Tl) showed a
slight shift towards slower water when the adults were
removed (Figure 21 and Table 13). In June 1990 (mean
juvenile size = 62.6 mm Tl.), the juveniles showed a
significant shift into faster water in response to adult
removal (Figure 22 and Table 10). The juveniles moved into
the 40 - 50 cm/sec velocity range but still avoided
velocities greater than 50 cm/sec. In August 1990 (mean
juvenile size = 66.5 mm Tl.), the juveniles responded to
adult removal by moving into the fastest velocity range
(Table 10 and Figure 23).

The juvenile habitat shift observed in the August 1990

riffle manipulation study was apparent in the diets of the

107
juveniles in the low adult treatment in that study (Table
15). These juveniles incorporated Simulidae in a fairly
high proportion into their diets (similar to the adults in
the high adult treatment) compared to the juveniles in the
high adult treatment. A similar change in diet was not
detected in the June 1990 study, even though there was a
significant habitat shift by the juveniles after adult
removal. This may be due to the fact that the juveniles
still avoided the fastest velocities in the absence of the
adults (Figure 22).

The faster areas within a riffle generally contain
higher invertebrate densities, especially of key longnose
dace diet items such as Simuliidae, Baetidae and
Hydropsychidae (Table 4, Brown and Brown 1984, Orth and
Maughan 1983). This higher food availability coupled with
the fact that the slower areas in a riffle are more likely
to contain the juveniles of many species of stream fishes
(Schlosser 1987, Schlosser 1988a, 1988b, Power 1987)
including; mottled sculpin (Cottus bairdi) and blacknose
dace (Rhinicthys atratulatus) in addition to adult and
juvenile burbot (Lota lota) (personal observation) may make
the faster areas of the riffle an energetically optimal
habitat for those fish that are able to use it. In order to
effectively utilize the faster areas, a fish must be able to
swim against the currents encountered there and maintain its

position behind or under a shelter. Even in the face of

108
possible competition for that shelter from other riffle
dwelling fish. Juvenile longnose dace are not able to swim
against the faster currents at small sizes (Figure 10). In
this study, the juveniles did not show a significant
response to adult removal until June 1990 when their mean
size was 62.6 mm Tl. Even then, they avoided velocities
greater than 50 cm/sec (against which they were not able to
swim (Figure 10)). By August of 1990, the juveniles had
reached a size (66.5 mm Tl) where they were able to utilize
the faster currents (Figure 10) and responded to adult
removal by moving into the fastest areas of the riffle
(Figure 23). However, in the presence of adults, in June
and August 1990, the juveniles used the slower velocities
(Table 10) suggesting that they were not able to compete
with the larger adults for space within the faster areas of
the riffles.

Throughout the course of this study, however, juvenile
longnose dace were captured in fast velocity areas
(sometimes in mean water column velocities exceeding 90
cm/sec). This velocity overlap with the adults indicates
that segregation between the adults and juveniles is only
partial. Partial segregation is likely to occur if the
faster habitat is the optimal habitat for both size classes
and shelter in that habitat is abundant enough to not be
completely limiting. Shelter from the current in fast areas

dominated by boulder and cobble substrates should be

109

abundant and, unless adult densities are very high, should
not be completely unavailable to the juveniles. In
addition, the velocity at the substrate surface is generally
lower than the mean water column velocity (Hynes 1970), and
although it may be prohibitively high in many areas, it
should be possible for juveniles to enter and utilize some
portion of the faster areas without actually encountering
prohibitive velocities. Juveniles in the faster areas,
would be able to take advantage of the higher invertebrate
densities in the faster areas because, with the exception of
Hydropsychidae, and the Simuliidae in early summer, most
invertebrates were as dense on the non—exposed surfaces of
rocks in the fast areas of riffles as they were on the
exposed surfaces (Table 16). However, a juvenile in a
faster area would be at risk of being displaced from its
shelter by a larger conspecific foraging in the area.

Ideally, evidence for competition includes some measure
of the effect of one group on the fitness of another
(Connell 1980). For fish, because of the strong
relationship between fecundity and size (Bagenal 1978),
growth may be used as one indicator of fitness (Werner et
al. 1983a). Several attempts were made to measure the
growth rates of juvenile longnose dace confined to slow and
fast areas within a riffle. However, due to the necessity
of using cages to confine the fish and the length of time

required to be able to determine growth rates, all attempts

110
were unsuccessful. Even with daily cleaning, the cages
affected the flow and invertebrate drift into the caged
area. In all cases, longnose dace confined to cages (% m2
in area) lost weight. In the absence of direct evidence of
a fitness effect, documented resource overlap between the
two groups and a reduction in the realized niche of one
group in the presence of the other is used as evidence for
competition (Colwell and Futuyma 1971, Diamond 1978, Gatz et
al. 1987). By the start of their second full summer of life
(June 1990), the juvenile longnose dace in this study showed
a high resource overlap with the adults (they both occurred
within riffles and their velocity ranges overlapped), and
they showed a reduction in their realized niche in the
presence of adults (as evidenced by the niche shift into
faster areas in the absence of adults) indicating that the
adults and juveniles competed for space in the faster areas
of the riffles.

In addition to providing higher invertebrate prey
densities for both juvenile and adult longnose dace, the
faster areas may also be important as a refuge from
predation. Adult longnose dace may reach the size were they
are vulnerable to avian predators such as kingfishers and
herons (both were seen frequently in the study area) and
terrestrial predators such as mink and raccoons (several
mink were seen in the study area). The faster areas within

a riffle are also the areas with the highest turbulence and

111

should, therefore, offer refuge from non-aquatic predators.
Also, the slower areas of a riffle are generally the
shallower areas of a riffle (Chapter one, Table 2) where
predation risk from avian and terrestrial predators is high
for larger fish (Power 1984, 1987, Schlosser 1988a). This
may explain the avoidance of shallow areas by adults
observed throughout this study. Smaller fish like juvenile
longnose dace are not as vulnerable to avian and terrestrial
predators (Power 1987, Schlosser 1988a) and, therefore,
would not need the refuge offered by the faster areas. This
lack of vulnerability, rather than physical constraints, may
explain the lack of a habitat shift into faster areas by the
juveniles after adult removal in June and August of 1989.
However, this study was not designed to examine habitat
specific predation risks, and this possibility is mentioned
only as a viable alternative to physical constraints as an
explanation of the juveniles habitat use in the 1989
studies.

Several studies document the effects of predation risk
on the habitat choice of fishes (Mittlebach 1981, 1984,
Werner et al. 1983b, Power 1984, Power at al. 1985,
Schlosser 1987, 1988a, Gotceitas and Colgan 1990). In all
these studies, small fish reduced their use of high risk
areas in the presence of predators, in some cases at the
cost of lowered foraging returns. To demonstrate that

predation risk in the slower shallower areas of a riffle

112
causes adult longnose dace to utilize the faster areas of
the riffle, it would be necessary to show that the adults
shifted their habitat use to the slower areas of the riffle
in the absence of predators. However, it may not be
possible to demonstrate a habitat shift in the absence of
predators. Habitat preferences produced by avian or
terrestrial predators may be innate. Predation risk has
been shown to cause innate behavior patterns in at least two
species of fish. Guppies (Poecilia reticulata) from areas
with high predation risk formed tighter schools and reacted
at a greater distance to risk when piscivorous fish were
present than did guppies from areas of low predation risk
(Seghers 1974). As these differences were maintained
through several generations of laboratory stocks, they are
most likely due to genetic differences between the two
groups. Three-spined sticklebacks (Gasterosteus aculeatus)
from areas with high predation risk develop an escape
response to heron attacks without prior experience to a
predator (Giles 1984). Members of this species taken from
an area with predators and an area without predators soon
after hatching were tested for their response to simulated
predators. Sticklebacks from the high predator area showed
a greater tendency to respond to the predator by either;
rapid jumping, freezing or taking cover in the weeds (when
available), than did the sticklebacks from the area without

predators (Tulley and Huntingford 1987). These responses

113
did not occur until the fish were at a size that is
vulnerable to predation (about 30 mm Tl). Innate predator
defense behaviors are likely to develop in situations where
the cost of learning (injury or mortality) to avoid
predators is fairly high. Innate behaviors probably develop
in stream fishes that are susceptable to avian and/or
terrestrial predators. Avian and terrestrial predators may
be difficult to detect for stream fishes, in which case,
anti-predator behaviors in response to predator detection
would be of little value, while avoidance of areas of
possible high predation risk (regardless of predator
presence or absence) would be a highly adaptive behavior.
This avoidance may be the case with large armored catfish
(Loricariidae) (Power 1984), bigmouth chub (Nocomis
platyrhynchus) (Lobb and Orth 1988), mottled sculpin (Cottus
bairdi) (Freeman and Stouder 1989) (see Mahon and Portt,
1985 for a summary of the size distribution of many species
of fishes occurring in riffles, raceways and pools) and the
larger longnose dace in this study.

Regardless of whether the faster areas of the riffle
are selected for their higher invertebrate densities or for
their value as a predation refuge (or both), there was a
consistent pattern of habitat segregation by velocity
between juvenile and adult longnose dace for at least the
first two full summers of the dace's life. This study

demonstrated that the cause of this pattern changed as the

114

juvenile longnose dace grew. At smaller sizes, even at low
adult densities, the juveniles did not use the faster areas
of the riffles. This avoidance of faster areas may be due
to physical constraints (as evidenced by the burst velocity
measurements) or due to a lack of predation risk to the
juveniles in the slower areas of the riffle. As the
juveniles grew, the pattern of velocity segregation was
maintained by intra-specific competition with the larger
longnose dace for space in the faster areas of the riffles.

This study was designed to assess the roles of physical
constraints and intra-specific competition in determining
the habitat use of the juvenile longnose dace. However, the
results from the 1989 studies and the fact that adult
densities do not explain the consistent pattern of substrate
segregation suggest longnose dace habitat use may be
affected by other factors in addition to physical

constraints and intra-specific competition.

CHAPTER FOUR

MORPHOMETRIC ANALYSIS

115

116

INTRODUCTION

As discussed above, the inability of juvenile longnose
dace less than 60 mm T1 to swim faster than 50 cm/sec may be
the initial cause of velocity segregation between the adults
and juveniles. Maximum swimming velocity is a function of
both body size and body shape (Moyle and Cech 1982). Fish
with elongate bodies, flattened heads, large forked caudal
fins and posteriorly placed dorsal and anal fins are highly
adapted for rapid swimming. Longnose dace may not develop
morphometric adaptations for faster burst swimming until
later in life, thus accounting for velocity segregation
between the size classes.

The purpose of this study was to examine the degree of
morphometric variation between, and within, the juvenile and

adult size groups.

METHODS

Data on the shapes of 87 longnose dace ranging in size
from 36 to 121 mm T1 were gathered according to the methods
provided by Bookstein et. al. (1985) for quantifying shape
change between groups of varying sizes. Longnose dace were
captured from the Ford river by kick seining and

photographed in late September 1987. Each dace was

117

anesthetized with MS 222 (Finquil), wiped dry and
photographed against a white background using a 28 - 80 mm
macro-zoom lens attached to a 35 mm camera mounted on a
tripod. Each photograph contained: the fish, an identifying
tag and a section of a ruler. The photographs were enlarged
and shape measurements were made by hand to the nearest 0.5
mm. The ruler in each photograph was used to convert all
measurements to actual size.

Six landmarks were chosen on the fishes profile and a
"box truss" (Bookstein et. al. 1985) was drawn connecting
all six landmarks (Figure 24). This allowed eleven shape
measurements on which principal component analysis (using
S.A.S. on the MSU IBM mainframe computer) was conducted to
examine the pattern of variation between three size groups
(36 - 56 mm T1, 65 - 80 mm T1, and 81 - 121 mm T1) and
within each size group. The division into size groups was
based on the length frequency distribution of the 87
longnose dace used in this study. Originally, the analysis
of shape variation within and between size groups was to be
conducted using the "shear" method of size free shape
discrimination (Bookstein et. al. 1985). However,
"shearing" does not necessarily remove size variation from
the shape analysis (Rohlf and Bookstein 1987). Hence,
principal component analysis was chosen as the more
effective means of examining the patterns of shape variation

in longnose dace.

118

 

Figure 24: Outline of an adult longnose dace showing the
landmarks and measurements used for the analysis
of shape change with growth.

119

RESULTS

Overall, over 97% of the variation in measurements can
be attributed to size variation within the data set (all
variables (factors) load significantly onto the first
principal component) (Table 17). An examination of the
individual size groups separately indicates that there is
some shape variation within each size groups. Within the
fry size group, over 84% of the variation is due to size and
7.4% of the shape variation (PC2) is related to the
insertion of the dorsal fin. In the juvenile size class,
70.4% of the variation is due to size, 14.6% (PC2) is
related to the insertion of the anal fin and 7.5% is due to
PC3 which is difficult to interpret but may be related to
the insertion of the dorsal fin. Most (84.1%) of the
variation within the adult size class is due to size, and
PC2 and PC3 (accounting for 11.9% of the variation) seem to

both relate to the placement of the dorsal and anal fins.

DISCUSSION

Principal component analysis of the eleven shape
variables for the 87 longnose dace in this study failed to
detect any significant variation in shape within the entire

data set. This may be because the large variation in size

120

Table 17: Results of principal component analysis on eleven
shape measurements from 87 longnose dace. Factors with
loadings greater than 0.2 (on a scale of 0 - 1) were
considered major factors for a component. Factors are
listed in order of decreasing loadings. The shape variables
designated by each letter are illustrated in figure 24.

 

 

Group Component % Variation Major Factors
ALL PCl 97.8 ALL (size)
FRY PCl 84.3 ALL (size)
N = 26 PC2 7.4 B,F,J,G,E
JUVENILES PCl 70.4 ALL (size)
N = 31 PC2 14.6 H,F,I,D,C
PC3 7.5 B,J,E,K,G,F
ADULTS PCl 84.1 ALL (size)
N = 30 PC2 6.8 F,H,B,I,E,G,J
PC3 5.1 H,I,G,J,E,B,C

 

121
(PC1) of the dace used swamped any variation in shape.
Indeed, when the size groups were considered separately,
more of the variation within each group appeared to be
related to shape, or, more specifically, to the placement of
the dorsal and anal fins.

This shape variation is relatively small and may not be
ecologically important to the fish. However, posterior
placement of the dorsal and anal fins is an adaptive trait
for fish that are lie in wait predators (Moyle and Cech
1982). Lie in wait predators must be capable of bursting
forward at rapid speeds in order to capture prey. Posterior
placement of the dorsal and anal fins places the trailing
edge (the part of the fin that delivers the propulsive
thrust) of those fins far back on the body where the
greatest amount of movement occurs during the tail beat,
thereby increasing the thrust derived from those fins.
Although longnose dace are not classic lie in wait
predators, their mode of existence in stream riffles may
require the same morphometric adaptations that are useful to
lie in wait predators. Longnose dace spend most of their
time under or behind shelters in the riffles. When
foraging, they will exit the shelter and burst forward into
the current to grab a prey item off of the substrate surface
and then drift back to the same or another shelter (personal
observation). Posterior placement of the dorsal and anal

fins would aid in their ability to burst forward and would

122
be adaptive to this mode of existence.
Because of the small amount of variation detected
between groups, however, there is insufficient evidence to
support speculation that size related velocity segregation

in longnose dace is a function of morphometric differences

between the adults and juveniles.

SYNOPSIS

123

124

Longnose dace undergo a gradual niche shift into
faster, deeper areas of riffles and larger substrates as
they grow. Or, more precisely, the niche of the longnose
dace gradually expands to include faster velocities, deeper
areas and larger substrates as it grows. As a result, the
juvenile niche is encompassed within the niche of the adult,
and partial habitat segregation within a riffle occurs
between the juvenile and adult longnose dace.

Adult and juvenile longnose dace were consistently
segregated by current velocity throughout this study. The
faster areas of a riffle have higher invertebrate densities
and probably offer shelter from predation for those dace
that are physically able to cope with the current velocities
found there. When they reach 55 - 65 mm T1, in the absence
of other fish, juvenile longnose dace show a preference for
the faster areas. At this size, their maximum burst
velocities are fast enough to allow them to use some portion
of the faster areas. Prior to that size, the juveniles show
a preference for the slower areas of the riffle, possibly
due to a physical inability to cope with the faster
currents. This preference for slower velocities results in
size specific velocity segregation within the riffle area.
Velocity segregation between the juveniles and adults,
however, occurs consistently, at least until the juveniles
reach an average size of 67 mm T1 (the average size of the

juveniles at the end of the riffle manipulation studies).

125
Once the juveniles become large enough to utilize the faster
areas of the riffle, velocity segregation is maintained by
intra-specific competition for space in the faster areas of
the riffle. Adult longnose dace are superior competitors
and displace juveniles from behind or under shelters,
exposing them to the current, and possibly to predation.
When adult densities were high, partial velocity segregation
occurs between the adults and juveniles, but when the adult
densities were lowered, the juveniles (> 60 mm Tl) moved
into the faster areas of the riffle.

Adult and juvenile longnose dace exhibited partial
depth segregation during this study. However, depth
segregation was inconsistent, occurring in only five of the
ten cases where size specific depth use was examined. Depth
segregation was not related to the size of the juveniles,
and juvenile depth use was not affected by adult density
manipulations. In all cases where depth segregation
occurred, the adults used deeper areas than did the
juveniles. Adult longnose dace may use the deeper areas of
the riffle as a shelter from terrestrial or avian predators,
or adults may use the deeper areas of the riffle because, in
some cases, the deeper areas are also the faster areas.

Partial substrate segregation occurred fairly
consistently throughout this study. In all cases, the
adults used larger substrates than did the juveniles.

However, manipulations of adult densities had no effect on

126
the substrate use of the juveniles. Boulders generally are
associated with lower invertebrate densities than cobble
substrates (Orth and Maughan 1983), indicating that factors
other than foraging returns are responsible for the adult
preference for larger substrates. Adult longnose dace may
use the larger substrates as a shelter from avian or
terrestrial predators, or, conversely, juvenile longnose
dace may avoid larger substrates because they might harbor
aquatic piscivores, to which they are vulnerable.

Several factors appear to be important in determining
the habitat use of longnose dace. Those factors differ
according to the size of the dace. The habitat choice of
larger dace appears to be a function of foraging returns and
predation risk, while the habitat choice of smaller longnose
dace appears to be a function of physical constraints,
intra—specific competition and, possibly, predation risk.
These factors all interact to produce a temporally
consistent pattern of size related habitat segregation
within a stream riffle.

Niche separation by stream fishes occurs through a
variety of mechanisms (Ross 1986). Inter-specific
competition is responsible for the spatial separation of:
riffle sculpin (Cottus quloses) and speckled dace
(Rhinichthys osculus) (Baltz et. al. 1982), rainbow trout

(Salmo gairdneri) and brown trout (Salmo trutta) (Gatz et.

 

al. 1987), and redline darters (Etheostoma rufilineatum) and

127
tennessee snubnose darters (Etheostoma simoterum) (Greenburg
1988). In all three examples, removal of the dominant
species results in a niche shift by the subordinate species.
Intra-specific competition between juvenile and adult
mottled sculpins (Cottus bairdi) results in size related
depth segregation (Freeman and Stouder 1989). Temporal
differences in feeding patterns results in niche separation
between juvenile and adult creek chub (Semotilus
atromaculatus) (Magnan and FitzGerald 1984), and between
torrentfish (Cheimarrichthys fostrei) and bluegilled bullies
(Gobiomorphus hubbsi) (Scrimgeour and Winterbourn 1987).
Size specific predation risk results in separation between
juvenile and adult loracariid catfish (Power 1984). All
these mechanisms, in addition to physical constraints, can
act singly, simultaneously, or sequentially to reduce inter-
specific and intra—specific niche overlap in stream fishes.
More research into the interactions of these mechanisms is
essential to a better understanding of the patterns of
habitat use by stream fishes.

There are still several questions to be answered
concerning the factors affecting the habitat choice of
longnose dace. For instance, in one control riffle, the
adults and juveniles were segregated by velocity and
substrate type in June 1989, while in August 1989, they were
segregated by depth but not by substrate type or velocity.

The availability of each of these habitat variables in this

128
riffle differed from that of the other two riffles used in
the riffle manipulation studies. This riffle was also
shallower and slower during the August study than it was
during the June study. It is unclear how spatial and
temporal habitat variability affect the habitat choice and

size specific interactions of longnose dace.

LITERATURE CITED

129

130

LITERATURE CITED

Abbott, J.C., R.L. Dunbrack and C.D. Orr. 1985. The
interaction of size and experience in dominance
relationships of juvenile steelhead trout (Salmo
gairdneri). Behaviour 92: 241-253.

Bagenal, T.B. 1978. Aspects of fish fecundity. Pages 75-101
in S.D. Gerking, (ed). Ecology of freshwater fish
production. J Wiley and sons, New York, New York USA.

Baltz, D.M., P.B. Moyle and N.J. Knight. 1982. Competitive
interactions between benthic stream fishes, riffle
sculpin, Cottus ulosus, and speckled dace, Rhinichthys
osculus. Can. J. Fish. Aquat. Sci. 39: 1502-1511.

Bartnik, V.G. 1970. Reproductive isolation between two
sympatric dace, Rhinichthys atratulus and g.
cataractae, in Manitoba. J. Fish. Res. Bd. Can. 27:
2125-2141.

Biette, R.M., and G.H. Geen. 1980. Growth of underyearling
salmon (Oncorhynchus nerka) under constant and cyclic
temperatures in relation to live zooplankton ration
size. Can. J. Fish. Aquat. Sci. 37: 203-210.

 

Bookstein, F.L., B.C. Chernoff, R.L. Elder, J.M. Humphries,
G.R. Smith, and R.E. Strauss. 1985. Morphometrics in
evolutionary biology. Academy of Natural Sciences of
Philadelphia Special publ. 15. 277 pp.

Brett, J.R. 1971. Energetic responses of salmon to
temperature. A study of some thermal relations in the
physiology and freshwater ecology of sockeye salmon
(Oncorhvnchus nerka). Am. Zoo. 11: 99-113.

Brown, V.A. and K.B. Brown. 1984. Distribution of insects
within riffles of streams. Freshwat. Invertebr. Biol.
3(1): 2-11.

Colwell, R.K. and D.J. Futumya. 1971. On the measurement of
niche breadth and overlap. Ecology 52: 567-576.

Connell, J.H. 1980. Diversity and the coevolution of
competitors, or the gohst of competition past. Oikos
35: 131-138.

Cummins, K W. 1962. An evaluation of some techniques for the
collection and analysis of benthic samples with special
emphasis on lotic waters. Am. Mid. Nat. 67: 477-504.

131

Diamond, J.M. 1978. Niche shifts and the rediscovery of
interspecific competition. Am. Sci. 66: 322-331.

Freeman, M.C. and D.J. Stouder. 1989. Intraspecific
interactions influence size specific depth distribution
in Cottus bairdi. Environ. Biol. of Fish. 24(3): 231-
236.

Gatz, A.J., M.J. Sale and J.M. Loar. 1987. Habitat shifts in
rainbow trout: competitive influences of brown trout.
Oecologia 74: 7-19.

Gee, J.H. 1968. Adjustment of buoyancy of longnose dace
(Rhinichthys cataractae) in relation to water velocity.
J. Fish. Res. Bd. Can. 25: 1485-1496.

Gee, J.H. 1972. Adaptive variation in swimbladder length and
volume in dace, genus Rhinichthvs. J. Fish. Res. Bd.
Can. 29: 119-127.

Gerald, J.W. 1966. Food habits of the longnose dace,
Rhinichthvs cataractae. Copeia 3: 478-485.

Gibbons, J.R.H. and J.H. Gee. 1972. Ecological segregation
between longnose and blacknose dace (genus Rhinichthys)
in the Mink River, Manitoba. J. Fish. Res. Bd. Can.
29: 1245-1252.

Giles, N. 1984. Development of the overhead fright response
in wild and predator-naive three-spined sticklebacks.
Anim. Behav. 32: 276-279.

Gotceitas, V. and P. Colgan. 1990. The effects of prey
availability and predation risk on habitat selection by
juvenile bluegill sunfish. Copeia 2: 409-417.

Greenberg, L.A. 1988. Interactive segregation between the
stream fishes Etheostome simoterum and E.
rufilineatum. Oikos 51: 193-202.

Hall, D.J., E.E. Werner, J.F. Gilliam, C.G. Mittlebach, D.
Howard, C.G. Doner, J.A. Dickerman, and A.J. Stewart.
1979. Diel foraging behavior and prey selection in the
golden shiner (Notemiqonug crvsoleucas). J. Fish. Res.
Bd. Can. 36: 1029-1039.

Helfman, G.S. 1986. Fish behavior by day, night, and
twilight. lg T.J. Pitcher (ed). The behavior of teleost
fishes. Croom-Helm, London. pp. 366-387.

132

Hynes, H.B.N. 1970. The ecology of running waters.
University of Toronto Press, Toronto, Ontario, Canada.
555 pp.

Jenkins, T.M. 1969. Social structure, position choice and
micro-distribution of two trout species (Salmo trutta
and Salmo gairdneri) resident in mountain streams.
Anim. Behav. Monogr. 2: 57 - 123.

 

Lee, D.S., C.R. Gilbert, G.H. Hocutt, R.E. Jenkins, D.E.
McAllister and J.R. Stauffer, Jr. 1980. Atlas of North
American freshwater fishes. North Carolina Biological
Survey Publ. 1980-12. 854 pp.

Lobb, M.D. III and D.J. Orth. 1988. Microhabitat use by the
bigmouth chub Nocomis platvrhvnchus in the New River,
West Virginia. Am. Mid. Nat. 120(1): 32-40.

Magnan, P. and G.J. FitzGerald. 1984. Ontogenetic changes in
diel activity, food habitats and spatial distribution
of juvenile and adult creek chub, Semotilus
atromaculatus. Environ. Biol. Fish. 11(4): 301-307.

Mahon, R. and C.B. Portt. 1985. Local size related
segregation of fishes in streams. Arch. Hydrobiol.
103(2): 267-271.

Meisner, D.J. 1990. Potential loss of thermal habitat for
brook trout, due to climatic warming, in two southern
Ontario streams. Trans. Am. Fish. Soc. 119: 282-291.

Mittlebach, G.G. 1981. Foraging efficiency and body size: a
study of optimal diet and habitat use by bleugills.
Ecology 62: 1370-1386.

Mittlebach, G.G. 1984. Predation and resource partitioning
in two sunfishes (Centrarchidae). Ecology 65: 499-513.

Moyle, P.B. and J.J. Cech Jr. 1982. Fishes: an introduction
to ichthyology. Prentice-Hall Inc. Englewood Cliffs,
N.J. USA 593 pp.

Orth, D.J. and O.E. Maughan. 1983. Microhabitat preferences
of benthic fauna in a woodland stream. Hydrobiol. 106:
157-168.

Pappantoniou, A. and G. Dale. 1982. Comparative food habits
of two minnow species: blacknose dace, Rhinichthys
atratulus and longnose dace, Rhinichthys cataractae.
J. Freshwat. Ecol. 1(4): 361-364.

133
1979.

 

Peterson, R.H., A.M. Sutterlin, and J.L. Metcalfe.
Temperature preference of several species of Salmo and

Salvelinus and some of theur hybrids. J. Fish Res. Bd.

Can. 36: 1137-1140.

Power, M.E. 1984. Depth distributions of armored catfish:
predator induced resource avoidance? Ecology 65: 523-
528.

Power, M.E. 1987. Predator avoidance by grazing fishes in
temperate and tropical streams: importance of stream
depth and prey size. In: Kerfoot, W.C. and Sih, A.
(eds), Predation: direct and indirect impacts on
aquatic communities. Univ. Press of New England,

Hanover, NH, pp. 333-353.

Power, M.E. and W.J. Mathews. 1983. Algae-grazing minnows
(Campostoma anomalum), piscivorous bass (Micropterus
spp.) and the distribution of attached algae in a small
prairie-margin stream. Oecol. (Berlin) 60: 328-332.

Power, M.E., W.J. Mathews, and A.J. Stewart. 1985. Grazing

piscivorous bass, and stream algae: dynamics

minnows,
of a strong interaction. Ecology 66: 1448-1456.

and D.J. McQueen. 1988. Ontogenetic changes in

Post, J.R.
the distribution of larval and juvenile yellow perch
(Perca flavescens): a response to prey or predators?
1820 - 1826.

Can. J. Fish. Aquat. Sci. 45:

Puckett, K.J. and L.M. Dill. 1985. The energetics of feeding
territoriality in juvenile coho salmon (Oncorhynchus

kisutch). Behaviour 92: 97-111.
1959. Age, growth, and food of the longnose dace,

 

Reed, R.J.
Rhinichthys cataractae, in northwestern Pennsylvania.

Copeia 160-162.
1987. A comment on shearing

Rohlf, F.J. and F.L. Bookstein.
Syst. 2001. 36: 356

as a method for "size correction".

- 367.
1986. Resource partitioning in fish assemblages:

Ross, S.T.
a review of field studies. Copeia 1986: 352-388.
Schlosser, I.J. 1987. The role of predation in age and size
related habitat use by stream fishes. Ecology 66: 651-
659.
Schlosser, I.J. 1988a. Predation risk and habitat selection
by two size classes of a stream cyprinid: experimental
test of a hypothesis. Oikos 52(1): 36-40.

134

Schlosser, I.J. 1988b. Predation rates and the behavioral
response of adult brassy minnows (Hybognathus
hankinsoni) to creek chub and smallmouth bass
predators. Copeia 3: 691-698.

Schoener, T.W. 1970. Nonsynchronous spatial overlap of
lizards in patchy habitats. Ecology 51(3): 408-418.

Scrimgeour, G.J. and M.J. Winterbourn. 1987. Diet, food
resource partitioning and feeding periodicity of two
riffle-dwelling fish species in a New Zealand river. J.
Fish Biol. 31: 309-324.

Seghers, B. 1974. Schlooing behavior in the guppy (Poecilia
reticulata): an evolutionary response to predators.
Oecologia 14: 93-98.

Sih, A. 1982. Foraging stratigies and avoidance of predation
by an aquatic insect, Notonecta hoffmanni. Ecology 63:
786-796.

Tulley, J.J. and F.A. Huntingford. 1987. Age, experience and
the development of adaptive variation in antipredator
responses in three-spined sticklebacks (Gasterosteus
aculeatus). Ethology 75: 285-290.

Vogel, S. 1988. Life's devices: the physical world of
animals and plants. Princeton University Press.
Princton, N.J. USA. 357 pp.

Werner, E.E. and J.F. Gilliam. 1984. The ontogenetic niche
and species interactions in size-structured
populations. Ann. Rev. Ecol. Syst. 15: 393-425.

Werner, E.E., G.G. Mittelbach, D.J. Hall and J.F. Gilliam.
1983a. Experimental tests of optimal habitat use in
fish: the role of relative habitat profitability.
Ecology 64: 1549-1555.

Werner, E.E., J.F. Gilliam, D.J. Hall and G.G. Mittelbach.
1983b. An experimental test of the effects of predation
risk on habitat use in fish. Ecology 64: 1540-1548.

Woodman, D.A. 1986. Microgeographic variation in Rhinichthys
cataractae (Cyprinidae) in western Nebraska. Copeia 1:
219-221.

Zippin, C. 1958. The removal method of population
estimation. J. Wild. Man. 22(1): 82-90.

Zar, J.H. 1984. Biostatistical analysis. Prentice-Hall Inc.
Englewood Cliffs, N.J. USA. 718 pp.

MICHIGAN STATE UNIV. LIBRARIES
MI WI l W ”I W II" II! W" “H llll ll) WW WI
31293009101928