..I I llrlulllulm ilwll- w

m
l

l/ Illl/l/ll/I

/

This is to certify that the

thesis entitled

EFFECTS OF LARGE-BODIED CLADOCERAN INDUCED INTERFERENCE
AND EXPLOITATIVE COMPETITION UPON ROTIFER COMMUNITY
STRUCTURE

presented by

Steven Charles Fradkin

has been accepted towards fulfillment
of the requirements for

M. S . Zoology

degree in

 

 

M ' rprofessor

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

LIBRARY
Michigan State
University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

_:l L

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l—jl ll

MSU Is An Affirmative Action/Equal Opportunity Institution
cMMoma-M

 

 

EFFECTS OF LARGE-BODIED CLADOCERAN INDUCED INTERFERENCE
AND EXPLOITATIVE COMPETITION UPON ROTIFER COMMUNITY
STRUCTURE
By

Steven Charles Fradkin

A THESIS

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

MASTER OF SCIENCE

Department of Zoology
Interdisciplinary program in Ecology and
Evolutionary Biology

1992

'\.

-531 ’ k

/"C/‘/ -— —

ABSTRACT

EFFECTS OF LARGE-BODIED CLADOCERAN INDUCED INTERFERENCE
AND EXPLOITATIVE COMPETITION UPON ROTIFER COMMUNITY
STRUCTURE

By

Steven Charles Fradkin

I examined the effects of a large-bodied cladoceran (Daphm'a pulicaria) upon a
natural rotifer assemblage to assess the relative magnitudes of interference and
exploitative competition. In situ bag enclosure experiments were conducted to test
hypotheses that rotifer populations of Anuraeopsis spp. , K. cochlearis typica, and
Polyarthra spp. are differentially suppressed via interference competition. Rotifers in
a short term experiment were exposed to a graded density of D. pulicaria. Rotifer
density decreases, observed vs. expected mortality rates, and egg ratio data were all
consistent with the hypothesis of suppression via interference competition. Some
suppression via food limitation caused by algal settling and exploitative competition
was inferred by decreased egg ratio absolute values over time. Rotifer observed
mortality rates were consistent with the hypothesis of differential suppression due to
differential interference susceptibilities. An unsuccessful long term experiment
supported the hypothesis of exaggerated algal settling food limitation in the short term

experiment.

ACKNOWLEDGMENTS

I would like to thank my major professor Dr. DJ. Hall, and the other
members of my guidance committee, Dr. A.J. Tessier, Dr. G.G. Mittlebach, and Dr.
S. Kalisz for their advice and support during this study. C.K. Geedey, Dr. M.
Leibold, Dr. A.J. Tessier, and Dr. C. Osenberg graciously provided helpful advice
and materials for the construction of bag enclosures. C.K. Geedey and A. Mullard
kindly provided a helping hand in the field. I would also like to thank W. Sobczak
and M. Wiplfi for many hours of enlightening discussion during the composition of
this thesis. M. Rondinelli, E. Beach, M. Olson, and A. Turner also provided useful
insights. W.C. Johnson kindly allowed unrestricted access to Wintergreen Lake. To
these people, I am truly grateful.

Financial support was graciously provided by a Department of Zoology
teaching assistantship and fellowship, an Ecology and Evolutionary Biology program
graduate fellowship, a W.K. Kellogg Biological Station teaching assistantship, and a
G.I-I Lauff Research award. I would also like to acknowledge Lisa A. Hallock for
making illustrations, providing unyielding support, and for putting up with me in the

field.

iii

TABLE OF CONTENTS

LIST OF TABLES

 

LIST OF FIGURES

 

CHAPTER I: Introduction

 

 

Study site:

CHAPTER 11: Short Term Experiment

 

METHODS

 

 

Sample collection and preservation

 

Sample processing

RESULTS

 

 

Rotifer suppression

 

Mortality rates: observed and expected

 

DISCUSSION

 

Rotifer suppression

 

CHAPTER 111: Long term experiment

METHODS

 

 

Sample collection and preservation

 

Sample processing

iv

vi

vii

17
18
21
23
26
27
35
57
59
67
68
69

7O

RESULTS AND DISCUSSION

 

CHAPTER IV: Conclusions and recommendations

Conclusions

 

 

Recommendations

 

BIBLIOGRAPHY

 

71

80

8O

83

87

TABLE

Table 1

Table 2

Table 3

Table 4

Table 5

Table 6

Table 7

LIST OF TABLES

Species list of rotifers found in Wintergreen lake
during the period of experimentation. * denotes common
species, ** denotes species that achieve high densities.

Initial and final stocking densities of Large Cladocerans
(Daphnia pulicaria) in experimental bags. Final densities
were used as working densities. Two bags were disqualified
from the analysis.

Coefficients of variation of between subsamples (within a
sample), and between samples. An asterisk (*) denotes data
not taken on species.

Paired t-test comparisons of mean change in observed daily
mortality rate from zero between three rotifer species. Tests
if mortality rates are different. An asterisk (*) denotes
significance at the a=0.05 level.

Wilcoxon Signed Ranks test comparing Observed vs.
Expected daily mortality rates within each species. An
asterisk C“) denotes significance at the a=0.05 level.

Paired t—test comparisons of mean change in egg ratio and
zero from zero. Tests if egg ratios are significantly different
from zero. Day = 5&6 is the lumped average of Day = 5
and Day = 6. An asterisk (*) denotes significance at the
a=0.05 level.

Long term experiment Large-bodied Cladoceran Densities
by Day and Treatment number.

vi

PAGE

16

25

28

4O

42

48

73

LIST OF FIGURES

FIGURE PAGE
Figure 1 Morphology of Daphnia large cladoceran entraining two

K. cochlearis typica into the branchial chamber. Note first

K. cochlean's typica is bearing an egg. 5

Figure 2 Bathymetric map of Wintergreen Lake, Michigan. Contour
intervals shown in meters. Location of experiments marked
by an asterisk C“) 10

Figure 3 Morphologies of rotifers examined in short and long term
experiments: Anuraeopsis spp. (Anur), K. cochlearis typica

(Kcoc), Polyarthra spp. (Poly) 11

Figure 4 Schematic representation of experimental design for the
short term bag enclosure experiment. 12

Figure 5 Schematic representation of experimental design for the
long term bag enclosure experiment. 13

Figure 6 Linear regression plot of weighted Anuraeopsis spp. density
by large cladoceran density for Day 5. Anuraeopsis spp.
densities are weighted by Day 0 densities. Y=-0.002X+0.119,
n=8, r2=0.581, F=8.314, P=0.028. Curved lines represent
95% confidence intervals. 29

vii

Figure 7 Linear regression plot of weighted Anuraeopsis spp. density
by large cladoceran density for Day 6. Polyarthra spp.
density weighted by Day 0 density. Y=-0.002X+O.112,
n=8, P=0.766, F=19.620, P=0.004. Curved lines represent
95 % confidence intervals.

Figure 8 Linear regression plot of weighted K. cochlearis typica
density by large cladoceran density for Day 5. K. cochlearis
typica density weighted by Day 0 density. Y=-0.004X+0.028,
n=8, r2=0.887, F=47.212, P=0.0001. Curved lines represent
95 % confidence intervals.

Figure 9 Linear regression plot of K. cochlearis typica density by
large cladoceran density for Day 6. K. cochlearis typica
density weighted by Day 0 density. Y=-0.001X+0.107,
n=8, r2=0.891, F=49.200, P=0.0001. Curved lines represent
95% confidence intervals.

Figure 10 Linear regression plot of weighted Polyarthra spp. density
by large cladoceran density for Day 5. Polyarthra spp.
density weighted by Day 0 density. Y=—0.003X+O.253,
n=8, r2=0.876, F=42.509, P=0.001. Curved lines represent
95% confidence intervals.

Figure 11 Linear regression plot of weighted Polyarthra spp. density
by large cladoceran density for Day 6. Polyarthra spp.
density weighted by Day 0 density. Y=-0.002X+O. 156,
n=8, r2=0.504, F=6.099, P=0.048. Curved lines represent
95 % confidence intervals.

Figure 12 Linear regression plot of Ceriodaphnia reticulara density
by large cladoceran density (treatment). Y=-3.006X+267.083,
n=8, P=O.644, F=10.877, P=0.016. Curved lines represent
95 % confidence intervals.

Figure 13 Linear regression plot of Bosmina longirosm's density by
large cladoceran density (treatment). Y=-O.895X+69.455,
n=8, r2=0.672, F=l2.274, P=0.013. Curved lines represent
95% confidence intervals.

viii

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

Observed daily mortality rates, Md, (% killed per day) for
Anuraeopsis spp. , K. cochlearis typica, and Polyarthra spp..
Rates based on Day 0-5. STARVE is the line for expected
daily mortality rates from Exploitative competition based on

Gilbert’s (1985) starvation experiment with K. cochlean’s tecta.

Anuraeopsis spp. Observed and Expected daily mortality
rates, Md, (% killed per day). Expected rates are based on
interference effects solely (Burns and Gilbert 1986b).
Observed rates calculated from Day O«5.

K. cochlean's typica Observed and Expected daily mortality
rates, Md (% killed per day). Expected rates are based on
interference effects solely (Burns and Gilbert 1986b).
Observed rates calculated from Day 0-5.

Polyarrhra spp. Observed and Expected daily mortality
rates, M, (% Killed per day). Expected rates are based on
interference effects solely (Burns and Gilbert 1986b).
Observed rates calculated from Day 0-5.

Plot of linear regression on Egg ratio for K. cochlearis typica
for Day 0. Y=-0.001X+0.284, n=5, P=0.022, F=0.066,
P=O.814. Curved lines represent 95% confidence intervals.

Plot of linear regression on egg ratio for K. cochlearis typica
Day 5. Y=-0.001X+0.036, n=5, F=0.072, F=O.231,
P=O.664. Curved lines represent 95 % confidence intervals.

Plot of linear regression on Egg ratio for K. cochlearis typica
for Day 6 Y=-0.002X+0.056, n=5, r2=0.358, F=l.672,
P=0.287. Curved lines represent 95% confidence intervals.

Plot of linear regression on Egg ratio for Anuraeopsis spp.
for Day 0. Y=-OOOX+O.365, n=5, F=0.000, F=0.001,
P=0.981. Curved lines represent 95% confidence intervals.

Plot of linear regression on Egg ratio for Anuraeopsis spp.

for Day 5. Y=-0.005X+0.116, n=5, €=O.305, F=l.317,
P=0.334. Curved lines represent 95% confidence intervals.

ix

45

46

49

50

51

52

53

Figure 23

Figure 24

Figure 25

Figure 26

Figure 27

Figure 28

Plot of linear regression on Egg ratio for Anuraeopsis spp.
for Day 6. Y=-0.006X+O.156, n=5, r2=0.279, F=1.610,
P=O.360. Curved lines represent 95% confidence intervals.

Plot of K. cochlearis typica Day 5 power analysis on Egg
ratio showing Observed (Y=-0.001X+0.036) and the largest
interesting detectable difference from Observed (Power,
Y=—0.002X+0.057). Power to detect (1-B)=0.035. Curved
lines represent 95 % confidence intervals.

Long term experiment Anuraeopsis spp. density (#lL) for
each Bag by day.

Long term experiment K. cochlearis typica density (#IL)
for each Bag by Day.

Long term experiment Polyarthra spp. densities (#IL) for
each Bag by day.

Long term experiment K. cochlearis typica mean egg ratios
for each Bag by Day.

54

56

75

76

77

78

CHAPTER I:Introduction

In natural aquatic systems, freshwater zooplankton communities are often
structured by a myriad of interactions within the zooplankton and between other
components of the aquatic community (fish, invertebrate predators, etc.). These
interactions may occur both within and across trophic levels (Lynch 1979). While the
effects of these interactions are recognized, the mechanisms which drive them are
variably understood. DeMott (1989) discusses four factors which control the
mechanisms that structure zooplankton communities; predation, changes in
environmental conditions, competition, and the interaction between them. The " top
down" effects of mechanisms such as size-selective predation and the cascading
dynamics of piscivore predation upon grazing zooplankton have received much
attention and their importance in structuring communities is well documented
(Carpenter et a1. 1985, McQueen et al. 1986). Changing environmental conditions are
also important in structuring both zooplankton and phytoplankton communities.
Seasonal changes in temperature, light, and nutrient availability can drastically affect
food availability and quality. These resource based ”bottom up" mechanisms, select
for zooplankton communities with differing food thresholds and food specializations
(DeMott 1989). In addition to resource limitation, environmental conditions such as
suspended clays (Kirk and Gilbert 1990) and high cyanobacterial densities (Gilbert

1990, Gilbert and Durand 1990), may interfere with zooplankton feeding. Reduction

of efficient resource utilization can affect community structure in ways similar to that
of resource limitation.

Competition is a factor affecting community structure by limiting availability
or utilization of resources via species interactions. Two major types of competition
are exploitative and interference competition. Exploitative competition can be defined
as the differential use of a shared, limited resource (Tilman 1982), while interference
competition is broadly defined by a variety of mechanisms such as allelopathy,
territoriality, encounter, and overgrowth competition (reviewed by Schoener 1983).
Exploitative and interference competition operate by different mechanisms, but both
may result in competitive exclusion of inferior competitors.

Three major components of freshwater zooplankton communities are
cladocerans (Crustacea), rotifers (Rotifers), and copepods (Crustacea). Competition
among and between these groups has been demonstrated to impact zooplankton
community structure. The strongest competitive interactions between these groups
occur between cladocerans and rotifers (DeMott 1989). Most copepods are
omnivorous or predatory and therefore do not overlap in diet the way cladocerans and
rotifers do. Cladocerans and planktonic rotifers overlap spatially and temporally in the
water column (Neill 1984), and both groups exhibit great overlap in food resources,
feeding mainly on phytoplankton in the sZOpm range (Gilbert 1985). Thus
cladoceran/rotifer competition can be expected. Cladoceran species may be
categorized into large (:12 mm) and small (51.2 mm) bodied groups with respect to

competitive abilities. This 1.2 mm size is a threshold for different interference

competition and size-selective predation effects. In natural systems, large cladocerans
usually dominate over small cladocerans in the absence of mechanisms (such as size-
selective planktivore predation) which competitively releases small cladocerans. Large
cladocerans have been shown to variably suppress and even exclude small cladocerans
in experimental studies (Lynch 1978, Kerfoot and DeMott 1980, Von Ende and
Dempsey 1981). Copepod dominated systems (without large or small cladocerans) are
much rarer and are usually the result of intense vertebrate and invertebrate predation
pressure upon cladocerans (Byron et al. 1986, Scavia et a1. 1988). The role of rotifers
in these systems has not been extensively studied.

Cladoceran/rotifer competitive interactions depend greatly upon cladoceran
body size. In natural systems there is an obvious pattern in the distribution and co-
occurrence of rotifers and cladocerans. Generally there is an inverse relationship
between the presence of large cladocerans and the abundance and diversity of rotifers
(reviewed by Gilbert 1988a). Systems dominated by large cladocerans often have
rotifer assemblages that are limited in diversity and abundance (Adalsteinsson 1979
Dabom et a1. 1978, Van Donk et al. 1990). Systems dominated by small cladocerans
usually have diverse, abundant rotifer assemblages (Gilbert 1988a).

Traditionally exploitative competition has been used to explain these trends in
cladoceran/rotifer abundances, ie. large cladocerans are more efficient exploiters of
shared, limited food resources and suppress small cladocerans and rotifers (Neill
1984, Gilbert 1985). Large cladoceran induced rotifer suppression via exploitative

competition has been demonstrated (MacIsaac and Gilbert 1991b, Gilbert 1988a).

Recently, interference competition has been proposed as a possible mechanism
affecting large cladoceran/rotifer interactions (Gilbert 1985, Gilbert 1988a). Gilbert
defines interference competition as incidental encounter damage (lethal and/or
sublethal) to rotifers via entrainment into the branchial chamber of large cladocerans
during filter feeding activity (Figure 1).

Large cladocerans are at least an order of magnitude larger than the largest
rotifer. Susceptibility of rotifers to interference damage is chiefly depended upon
residence time in the branchial chamber. Only cladocerans above the body size of 1.2
mm have chambers large enough to accommodate rotifers, thus 1.2 mm is the body
size threshold for interference competition. Experimental work has demonstrated the
operation of interference competition under laboratory conditions (Gilbert and
Stemberger 1985, Burns and Gilbert 1986b, Schneider 1990, MacIsaac and Gilbert
1991b). Burns and Gilbert (1986a) made direct observations of interference
competition by the large cladoceran Daphnia magna on K. cochlearis tecta. Rotifer
entry into and residence time in large cladoceran branchial chambers and residence
time is associated with rotifer escape response, body size, and integument texture.
Planktonic rotifers are a morphologically diverse group of organisms, with a range of
sizes (70-400 um), integument types (strong loricae, weak loricae, no loricae), and
escape responses (Gilbert and Kirk 1988, Kirk and Gilbert 1988). Based on these
characteristics, rotifer species may be expected to exhibit differential susceptibility to
large cladoceran induced interference competition. Gilbert (1988b) has shown this to

be true in the laboratory for 10 rotifer species.

 

Figure l Morphology of Daphnia large cladoceran entraining two K. cochlearis

typica into the branchial chamber. Note first K. cochlean's typica is bearing an egg.

Both interference competition and exploitative competition can potentially
impact rotifer assemblages, yet as previously mentioned each operates by a different
mechanism. Both interference competition and exploitative competition are density
dependent, increasing in effect with large cladoceran density. When considered
separately in the short term, interference and exploitative competition are expected to
have different suppressive effects upon rotifer assemblages.

When interference competition is reasonably strong, rotifers can be expected to
be suppressed numerically. As a result, the time response of rotifer suppression
should be directly proportional to the volume of water filtered by the large cladoceran
community as determined by large cladoceran density. If food resources are high,
then rotifer birth rates should remain relatively high since birth rates are directly
related to food resources (Stemberger and Gilbert 1985). Large cladoceran induced
mechanical interference does have the potential for shipping eggs from subitaneously
egg carrying genera such as Anuraeopsis and Keratella. Separated eggs may still be
viable, but they may settle into less advantageous areas of the water column where
development and survivorship may be adversely affected (Gilbert 1988a).

Rotifer assemblages experiencing predominantly high exploitative competition
can be expected to exhibit numerical decreases caused by starvation and senescence.
Exploitative competition mainly impacts rotifer assemblages by affecting their birth
rates, thus suppressing rotifer recruitment. Rotifers exhibit very high rates of
population increases (Gilbert 1988a), and can be expected to go through several

generations within a period of several days.

Interference and exploitative competition do not occur separately in large
cladoceran/rotifer systems though. The above predictions can still be expected
depending upon the relative magnitudes of interference and exploitative competition in
a given system. Interference and exploitative effects are often confounded at very high
large cladoceran densities because high large cladoceran densities can quickly reduce
food resources and exhibit high rates of mechanical interference concurrently. As
previously mentioned, birth rates can provide insight into this quandary as an
indicator of resource levels. Interference and exploitative effects are also often
confounded at low densities of large cladocerans, because the magnitudes of
interference and exploitative competition between large cladocerans and rotifers may
be overshadowed by interspecifc competition between rotifers.

Gilbert and Stemberger (1985) identified four conditions upon which the
population effects of interference and exploitative competition relative magnitudes
depend. They are; 1) large cladoceran density, 2) rotifer density, 3) rotifer
susceptibility to interference competition and exploitative competition, and 4) food
availability.

The effect of large cladoceran density upon the relative importance of
interference and exploitative competition has already been largely discussed. High
rotifer densities increase both demand on limiting resources, and the probability of a
large cladoceran encountering a rotifer.

Rotifer susceptibility to exploitative competition is dependent both upon

a particular species’ ability to survive and reproduce at low food levels, and its level

of specialization of food types. There is significant variability in planktonic rotifers to
both limiting food levels and food specialization (Bogdan et al. 1980, Stemberger and
Gilbert 1985). Rotifer susceptibility to interference competition is dependant upon
morphological and behavioral specializations which affect residence time in the large
cladoceran branchial chamber. Rotifer species relatively unsusceptible to exploitative
competition are not necessarily the same species that would be expected to be
unsusceptible to interference competition. This is especially true if one assumes a
metabolic cost to features that impart interference competition resistance (heavy
lorica, spines, swift escape responses). Given a rotifer assemblage with differential
exploitative and interference competition susceptibilities, one may expect that in a
system dominated by interference competition that the rotifer species composition
could be shifted towards species less susceptible to interference competition. In
systems dominated by exploitative competition, however, one would expect selection
for species less susceptible to exploitative competition. Food availability greatly
impacts exploitative competition effects. If food is not limiting, then exploitative
competition is obviously greatly reduced.

Most field studies to date have confounded interference and exploitative
competition by failing to convincingly differentiate between the two. A notable
exception is a recent study by Wickham and Gilbert (1991). In their study they report
depressed population growth rates for susceptible rotifer species (K. cochlearis and K.
earlineae) and the persistence of an unsusceptible species (Polyarthra vulgan’s) in the

presence of a large cladoceran (D. pulex). Suppression was achieved at very low D.

pulex densities (s 1.0 inds-L“) with size ranges from 1.17-1.57 mm. Interference
effects at such a low large cladoceran density are expected to be very small (Burns
and Gilbert 1986b). Interference competition was inferred because food resources in
the experiment were abundant throughout.

The purpose of my study was to attempt to differentiate between and assess the
relative magnitudes of interference and exploitative competition under near natural
field conditions. To accomplish this goal, I conducted a short term (6 days) and a
long term (24 days) bag enclosure experiment in the limnetic zone of Wintergreen

Iake, Michigan, during the summer of 1991 (Figures 4 and 5).

Study site

Wintergreen Lake in Kalamazoo Co., southwest Michigan is a small (14.6 ha),
hypereutrophic, hardwater lake approximately 6.5 M maximum depth with a mean
depth of 3.5 M (Figure 2). It is located on the W.K. Kellogg Bird Sanctuary whose
thousands of migrant and resident waterfowl provide nutrients in the form of feces to
support its > 1000 mg C -M'2-day‘l (Wetzel 1983) annual mean primary productivity.
Additionally, nutrients enter the lake via agricultural runoff from nearby cultivated
fields (Hall and Ehlinger 1989, Manny et al., 1975).

Wintergreen Lake’s phytoplankton is dominated by small flagellated algae

(~1-18 pm) in the spring, but shifts to dominance by cyanobacteria in the

10

 

 

 

 

 

0 50 100 150m
AREA 15ha L__ 1 A 1

Figure 2 Bathymetric map of Wintergreen Lake, Michigan. Contour intervals shown
in meters. Location of experiments marked by an asterix (*)

ll

 

Anur

 

KCOC

100 um

Figure 3 Morphologies of rotifers examined in short and long term experiments:
Anuraeopsis spp. (Anur), K. cochlearis typica (Kcoc), Polyarrhra spp. (Poly).

12

2:885 was ECB team 2: c8 cwmmou 5:02:5on .8 5255852 ocmegom v 82w:—

 

aOI UZ< r1

«~32.

All

 

 

 

 

 

 

 

 

 

.Le
3C
:1:
j:
{11:

 

 

11;:

0(-

 

 

L

L

 

2050a wx(d

 

 

 

 

 

xu_¢o

 

 

L

 

 

E13:

H

31.3 :5

13

.EoEtomxo
8:865 web EB Ex: 05 c8 swede 358598 Lo cozscaoafl ocannom m 83E

 

GOIUZ<

 

 

 

 

 

 

 

«whwfip
.lliiillL

 

 

 

 

0<m

 

 

 

20.5.00 mx<4

 

 

 

XUEm

 

 

 

 

_ 1

 

 

 

 

 

 

Trill!

l~ w0<uCDw
wx<q

summer once nitrogen is reduced to limiting levels by non Nz-fixing phytoplankton
(Manny, 1971, Wetzel, 1983). The lake quickly becomes anoxic below 2.5 M in the
late spring once it becomes thermally stratified (Fradkin, pers. obs.).

Wintergreen Lake’s history makes it an interesting lake in which to conduct
in situ experiments on large cladoceran/rotifer competitive interactions because it has
experienced a shift from large cladoceran dominance to small cladoceran dominance.
In 1977 and 1978, the lake experienced severe winterkill events. In its pre-winterkill
incarnation, the lake’s fish community was dominated by the piscivorous largemouth
bass (Microptera salmoides) and the planktivorous bluegill sunfish (Lepomis
macrochirus). Low dissolved oxygen levels during the winterkill winters eliminated
these fish from the lake. Fish species able to withstand low dissolved oxygen levels
such as yellow perch (Perca flavescens), pumpkinseed sunfish (Lepomis gibbosus),
yellow bullhead (Ictalurus natalis), bowfm (Amia calva), and golden shiners
(Notemigonous crysoleucas) came to dominate the lake. Golden shiners in particular
flourished due to predation release from largemouth bass (Hall and Ehlinger 1989).

Coinciding with this shift in fish community structure, the zooplankton
community structure also changed. In its pre-winterkill condition, Wintergreen Lake
was dominated by a succession of large cladocerans Daphnia pulicaria and Daphnia
galeata mendotae. These large-bodied cladocerans at times achieved densities of up to
200 inds L" (Threlkeld 1977). Unfortunately, there is no strong data on the pre-
winterkill rotifer community, but given previously cited experimental results on the

nature of interference competition, it seems reasonable to assume that sustained large

14

cladoceran densities of 200 inds L" could have both an exploitative and interference
effect upon rotifers. If rotifers in the field exhibit differential interference
susceptibility, the pre-winterkill community could be expected to be sparse, consisting
of relatively unsusceptible species. Qualitative observations during the 1970’s support
this contention (Hall, unpub. data). In its post-winterkill condition, the lake was
depleted of large cladocerans by the greatly enhanced golden shiner population.
Wintergreen Lake became dominated by small cladocerans which were competitively
released (mainly Bosmina longirostris, Ceriodaphnia reticulum, Daphnia parvula, and
Diaphanosoma brachyurum, all < 1.3 mm) (Hall and Ehlinger 1989). The post-
winterkill rotifer community is diverse (Table l), and at times reaches levels of
6,000-10,000 inds. L" (Fradkin, pers. obs.).

Thus I hypothesize the pre and post-winterkill rotifer communities to be
significantly different in Wintergreen Lake, both in types of rotifers present and in
abundance. I furthermore hypothesize that due to different expected effects of
interference and exploitative competition, differences in rotifer community structure
should be directly influenced by the relative roles of exploitative and interference
competition as mediated by large cladocerans. As a site for in situ experiments,
Wintergreen Lake is an appropriate system with a diverse, abundant rotifer

community that has been free from interference effects for 13 years.

15

16

Table 1 Species list of rotifers found in Wintergreen lake during the period Of
experimentation. * denotes common species, ** denotes species that numerically
dominant species.

 

Species

 

Anuraeopsis fissa **

 

A. navicula **

 

Brachionus angularis

 

B. budapestinensis
B. caudatus

B. Quadridentatus
Cephalodella gibba
Collotheca mutabilis

 

 

 

 

 

Colurella uncinara

 

 

 

 

Conochiloides dossuarius
Euchlanis dilatata

F ilinia longiseta *
Fllinia tenninalr’s *

 

Keratella cochleart's typica **

 

K. cochlearis tecta **

 

K. quadrata

 

Lecane luna

Lepadella patella

 

 

Monastyla closterocerca

 

 

 

 

 

 

M. lunaris

Palyarthra vulgaris **
P. remara **
P. major **
Synchaeta 3p.

Mchocerca similis *

 

 

 

T. pusilla '1‘

 

Chapter 11: Short Term Experiment

The purpose of this experiment was to assess the short term relative effects of
interference and exploitative competition by large-bodied cladocerans (large
cladocerans) upon rotifer community structure. In particular, the answers to three
questions were sought; 1) Does the presence of introduced large cladocerans into a
natural system dominated by rotifers suppress rotifer populations?, 2) If manifested,
does suppression differentially affect rotifer species?, 3) If manifested, is suppression
more consistent with expected effects of interference (physical encounter) competition
than those of exploitatative (resource) competition? The null hypotheses tested were;
1) large cladoceran introduction has no effect upon rotifer population growth, 2) large
cladoceran suppression affects all rotifer species the same, and 3) large cladoceran
suppression is indistinguishable from predicted exploitative competition effects.

For this study, three rotifer genera were selected for examining population
dynamics; Anuraeopsis spp. , Keratella cochlearis typica, and Polyarthra spp (Figure
3). These species were chosen because they represent the range of predicted relative
interference competition susceptibilities (high, medium, low respectively) (Gilbert,
Pers. comm., Gilbert 1988b), and because they were numerically abundant in
Wintergreen Lake. Anuraeopsis spp. possess a weak lorica (mean lorica length
IOOuM) with no posterior spine. Keratella cochlearis typica is the ubiquitous spined

morph with a mean lorica length of 162 ”M. Its lorica is somewhat more robust than

17

Anuraeopsis’s. Polyan‘hra spp. are illoricate rotifers (mean body length 124 uM) with
strong, paired swimming fins (paddles) enabling swift escape responses (Kirk et al. ,
1988, Stemberger 1979).

Anuraeopsis and Kerarella produce easily countable eggs for calculation of egg
ratios, which as an index of reproduction are an indirect measure of food availability.
Egg ratios are defined by Edmondson (1965) as the ratio of eggs/female. Stemberger
and Gilbert (1985) and MacIsaac and Gilbert (1991b) demonstrate in the laboratory
that there is a tight linkage between food level and egg ratio in rotifers. Rotifers
allocate much of their energy to reproduction and can exhibit maximal reproductive
rates (rm) of 0.3—0.8 d", largely dependent upon temperature. The eggs of Polyarthra
spp. are not easily counted, because they are not easily differentiated from the body
of the organism.

METHODS

On 7/11/91 a six day short-term experiment was initiated into Wintergreen
Lake (Figure 2) consisting of ten bag enclosures (Figure 4). Each enclosure housed a
natural Wintergreen Iake plankton community which included a diverse, abundant
rotifer assemblage (the above three rotifer genera and the lesser abundant species in
Table 1), phytoplankton, and a natural small cladoceran assemblage (predominantly
Bosmina longirostris, and Ceriodaphnia reticulata). The experiment was designed to
have five treatments with two replicates per treatment. Treatment bags were
inoculated with large cladocerans to produce a gradient of large cladoceran densities

(0,0.5, 5.0, 50, and 200 inds-L" respectively).

18

Bag enclosures were constructed from 4 mil, 20 cm diameter, 2.5 M deep
polyethylene tubing, holding approximately 72.6 liters. Bags were open at the top and
fastened to a wooden embroidery hoop that rested in a wooden frame fitted with a
styrofoam collar for floatation. An initial trial of this experiment (6/11/91) was foiled
by massive additions of fecal matter from waterfowl that perched on the enclosures.
Therefore, a piece of plastic sheeting was sandwiched between the embroidery hoop
and bag material, forming a 15 cm tall barrier to exclude bird droppings from the
bag. Bag bottoms were heat-sealed and cable-tied to stop leakage. A brick was cable-
tied to excess material at the bag bottom to make it negatively buoyant enough to
remain in a vertical position. The resultant apparatus produced a column of water 2.5
M deep exposed to natural, variable environmental conditions. Bags were tethered in
a series in line with the normal wind direction to reduce the likelihood that the
experimental set-up would break loose in the lake.

Between 1300-1600 hrs on 7/11/91 (Day minus one Of experiment) bag
enclosures were stocked with Wintergreen Lake water and its resident plankton
community. Bags were filled by hand via 20 liter pails from upper meter of
Wintergreen Lake where the bulk of the rotiferan and cladoceran community resides
due to anoxic conditions below 2.5 M (Fradkin, pers. obs.). A 7 cm head of water
was established in each bag to apply internal bag pressure enabling it to maintain its
cylindrical form.

For this experiment, Daphnia pulicaria was the large cladoceran chosen for

stocking because 1) its adult body length (top of head to base of spine) is well above

19

the predicted 1.2 mm size threshold for interference competition (Burns and Gilbert
1986b, Gilbert 1988a), 2) it is a hearty daphnid that usually avoids surface tension
entrapment (A.J. Tessier pers. comm.), 3) the species had been a dominant member
of the lake’s pre-winterkill zooplankton community, and 4) it was locally abundant in
nearby lakes at the time Of the experiment. Other cladocerans in Wintergreen Lake on
occasion may breach the 1.2 mm body size threshold, especially Ceriodaphnia. NO
Ceriodaphnia over 1.2 mm were recorded for this experiment though some were in
the long-term experiment, (see Chapter 111).

On 7/12/91 (Day zero of the experiment) D. pulicaria were collected between
0900-1000 hrs. in Gull Lake (approximately 2.5 km from Wintergreen Lake). Large
D. pulicaria were collected with a 1 M diameter 505 um mesh vertical haul net from
a depth of 15-20 M (D. pulicaria were most concentrated at this depth (S. Hu, pers
comm.)). A 505 pm mesh net was used to capture only the largest individuals
ensuring stocking animal above 1.2 mm. Stocking with only large adults was intended
to permit a relatively stable large cladoceran density over the course of the 6 day
experiment. Any D. pulicaria born during the six days were not expected to grow
above the 1.2 mm size threshold in that time span. The only factor expected to alter
large cladoceran density was mortality of the stocked individuals. The netted large
cladocerans were concentrated in the plankton bucket of the net and transferred to a
20 liter pail containing fresh Gull Lake water. The large cladocerans were then
transported to the Kellogg Biological Station (on Gull Lake) where the pail’s contents

were mixed haphazardly and 3 l-liter aliquots were collected. Aliquots were then

20

preserved and counted in tom, and body lengths (top of head to base of spine) of the
first 30 individuals encountered in each aliquot were recorded. The number of
animals in 100 mls of randomly mixed stocking pail water was computed along with
the appropriate volume to be added to each treatment (ie. Volume X large cladoceran
density = expected stocking density). During the counting/calculation process, the
large cladoceran pail was stored in a walk-in refrigerator to lower the water
temperature, reducing oxygen consumption by the animals in the pail (M. Leibold,
pers. comm.). The large cladocerans were then transported to the field where
treatment bags were inoculated. The entire collection/counting/transport process took
~ 2.5 hrs. Assignment of bags to treatment conditions was derived from a random

number table (Gill 1978).

Sample Collection and Preservation

Rotifers were sampled on days 0, 5 and 6 using a sampler made from a 2.54
cm diameter, 2.5 M long piece of clear plastic tubing holding a volume of 1.27 L.
The tube was weighted near the bottom (via duct taped metal weights) to keep the
tube straight during descent. Sampling was conducted by lowering the tube to the
bottom of the 2.5 M bag, corking the top, and smoothly withdrawing the tube. The
tube end was then immediately placed into a 2 liter pail, and gently uncorked allowing
the vertically integrated sample to flow into the pail. The sample was then

concentrated by pouring it through a 45 pm nitex screen, transferred to a labelled 100

21

ml QuorpakTM bottle, and stored on ice. Two replicate samples were taken from each
bag. Samples were preserved in the laboratory after sampling using a 5% cold sugar
formalin solution (Prepas 1978). Samples were stored on ice in the field to reduce
activity in the concentrated samples. While the sampler is inadequate for sampling
cladocerans, it seemed adequate for rotifers because of their small size (70-200 pm)
and minor escape responses relative to the diameter of the sampler.

Aside from aliquot sampling to determine initial stocking density, the large
cladocerans were sampled on day 6 at the close of the experiment. Small enclosure
size enabled the entire bag contents to be passed through a 150 pm plankton net to
collect all cladocerans (both large cladocerans and smaller species). These were
concentrated, transferred to labelled 150 ml Quorpakm bottles, stored on ice, and
preserved with 5 % cold sugar formalin in the laboratory. Small cladoceran species
were only sampled once, at the close of the experiment (Day 6).

Algal samples were collected using a 1.27 cm diameter, 2.5 M long hose.
Approximately 250 mls of vertically integrated bag water was collected into a
QuorpakTM bottle and preserved with an acid Lugol’s solution. Algal samples were
intended to provide data on relative availability of edible algae. Algal samples were
not counted, and thus were not used in the analysis of this experiment.

Temperature and 02 measurements were taken every meter in 3 selected bags
on Day 6. Data from a previous preliminary trial (6/21/91) indicated that temperature
and 02 profiles were not greatly different between bags. 02 measures were conducted

using a YSI model 57 oxygen meter, while a YSI model 43TD thermometer was used

22

for temperature measures.

Sample Processing

Rotifer samples were concentrated with a 40 pm screen and brought to a
constant volume of 20 ml. Subsamples of 1 ml were taken from the randomly mixed
sample with a wide bore pipet, and placed in a 1 ml Sedgwick-Rafter cell (Lind 1979,
Wetzel and Likens 1991). The entire cell was examined under 100x magnification,
and the abundance and type of rotifers along with the number of eggs still attached to
subitaneously egg-bearing genera (Anuraeopsis, Keratella) were counted. Repeated
subsamples were usually counted from each sample until at least 25 individuals from
each major rotifer group (Anuraeopsis spp. , Keratella cochlearis typica, and
Polyarthra spp.) were counted. At least 2 subsamples were always counted (ie. when
rotifer densities per subsample were high). Only 2 subsamples were counted in
samples that contained extremely low rotifer densities (ie. in the 40.8, 75.5, and 89.3
inds - L" large cladoceran treaments), consequently 25 individuals were not always
counted here. The presence of large blue green algal filaments in the samples
precluded counting of these samples in toto.

Large cladoceran samples were filtered through a 255 um screen, concentrated
and either counted in toto (for low density samples) or via 2.9 ml sampling buckets (a
small cup attached to a rod that is quickly dipped into the sample to extract 2.9 mls of

the mixed sample) drawn from a randomly mixed standardized 100 ml volume

23

containing the large cladocerans. Repeated aliquots were counted in higher density
treatments until at least 60 individuals were counted. The filtered sample was
examined for any large cladocerans that may have passed through the screen. Body
sizes for the first 30 large cladocerans encountered were measured using an optical
micrometer at 3.2x magnification to determine average body sizes for each treatment.
All animals were returned to the original filtrate for a subsequent analysis of small
cladocerans.

Small cladocerans and copepods (< 1.22 mm) were concentrated using a 90
pm screen, and brought to a standard volume of 100 mls. This was randomly mixed
and repeatedly subsampled (1 ml) using a wide-bore pipet until at least 50 of the 2
dominant (or most dominant if the other was scarce) small cladocerans were counted.
At least 2 subsamples were always counted. At least 20 of the major and 10 of the
minor small cladoceran species’ body sizes were measured to determine average small
cladoceran body size per treatment.

Water transparency measurements using a Secchi disk were taken for each
representative treatment on 7/ 18/91. Due to the small diameter of enclosures
compared to that of the Secchi disk, the disk was lowered into the lake beside the
transparent bags and viewed by looking down into the bag and through the side at the
Secchi disk. This procedure was meant to give some indication of algal content of the
bags relative to the lake. During the course of the experiment bags were not stirred.

Data from this experiment were analyzed largely by linear regression because

true replication was not achieved (Table 2). As a result, the geometric means of the

24

25

Table 2 Initial stocking, final, and Effective treatment densities of Large Cladocerans
(Daphnia pulicaria) in experimental bags. Effective Treatment densities were
calculated as geometric means of initial and final densities. Two bags were
disqualified from the analysis.

 

 

 

 

 

 

 

 

 

 

 

 

r- Initial Final Density Effective Comments
Stocking (#l L) Treatment
Density (#IL) Density (#IL)
0 Q Q Bag lost
0 0 0
5 1.9 .97
.5 2.5 1.1 Disqualified
5 1.8 3.0
5 5.0 5 .0
50 16.4 28.6
50 33.4 40.8
200 28.5 75.5
200 39.9 89.3 1

 

 

 

 

 

initial and final large cladoceran densities were used as effective treatment large
cladoceran densities during analysis. Rotifer egg ratios and densities were examined
using linear regression methods. Hypotheses concerning differences amongst
populations were examined using matched-paired t-tests or Wilcoxon Signed Ranks
tests, depending upon whether the data met the appropriate assumptions. Power
analyses were conducted on suspect linear regressions in which the null hypothesis
could not be rejected (Gerrodette 1987). Assumptions (normal distribution of errors,
constant variance of errors, independence of errors, etc.) were tested graphically

(Wilkinson 1990) for all statistical tests, and were met except where otherwise noted.

RESULTS

During the course of the experiment, one bag enclosure was lost due to failure
of the apparatus to secure the bag properly in the frame. Another was eventually
disqualified from analysis because; 1) it showed much higher rotifer densities than
treatments with similar large cladoceran densities, 2) values from this bag were
consistently identified as extreme outliers in every test in which they were used, and
3) the bag was visibly greener and had more periphyton growth on its side in relation
to similar treatments.

Table 2 shows the initial stocking (0-200 inds-L"), final (0-40 inds-L"), and
effective treatment densities of large cladocerans by bag. All bags had large

Cladocerans above the 1.2 mm size threshold, and the average body size for all bags

26

was 1.711024 mm. Coefficients of variation for sampling differences varied from
O.338-0.471 for between-sample rotifer abundances and 0343-0475 for between-
sample egg counts. Between-subsample coefficients of variation varied from 0.29-0.34
for rotifer abundances and 0.41-0.62 for egg counts (Fable 3).

Secchi depths were taken from the lake (0.7 M) and five bags with large
cladoceran densities of 0, 1.8, 1.9, 28.5, and 33.5 inds-L". Bag secchi depths were
0.8, 0.8, 0.8, 1.3, and 1.6 M respectively.

02 concentrations in the bags ranged from l2.0-6.6 mg -L" from surface to the
bottom based on 3 bags. 02 in the lake was 8.8 mg-L" at 2 M depth, and 0.3 mg-L"
at 3 M depth. Temperatures ranged from 28-24.8° C from bag surface to bottom.

These data are similar to temperatures found in the lake.

Rotifer Suppression

Figures 611 illustrate linear regressions for log weighted densities of
Anuraeopsis spp. , K. cochlearis typica, and Polyarthra spp. on Days 5 and 6 by large
cladoceran density (treatments). Log weighted densities were derived by subtracting
the log of the initial rotifer density from the log of the final rotifer density. Rotifer
densities were log weighted to compensate for between bag initial variation in rotifer
numbers due to stocking differences (Anuraeopsis spp. = 917314034 inds-L", K.
COChlearis typica = 718.0;t198.9 inds-L", and Polyarthra spp. = 21651663

inds-L" for Day 0). The log weighted density variable is thus related to the

27

28

Table 3 Coefficients of variation of between subsamples (within a sample), and
between samples. An asterisk (*) denotes data not taken on species.

 

Between subsample Between samples

 

 

Anuraeopsis spp. 0.34 0.41 0.471 0.343
Keratella cochlearis typica 0.28 0.62 0.368 0.475
Polyarthra spp. 0.29 * 0.338 * Jl

 

 

 

 

 

 

 

 

29

 

O3 7 I I I

0.2 —

Log Weighted Density

 

 

 

 

-01 l 1 l l
O 19 38 57 76 95

Large Cladoceran Density (it/L)

Figure 6 Linear regression plot of log weighted Anuraeopsis spp. density by large
Cladoceran density for Day 5. Anuraeopsis spp. densities are weighted by Day 0
densities. Y=-0.002X+0.119, n=8, r2=0.581, F=8.314, P=0.028. Curved lines
represent 95 % confidence intervals.

30

 

03 T l I I

0.2 - —

Log Weighted Density

 

 

 

 

O 19 88 57 76 95
Large Cladoceran Density (it/L)

Figure 7 Linear regression plot of log weighted Anuraeopsis spp. density by large
cladoceran density for Day 6. Polyarrhra spp. density weighted by Day 0 density.
Y=-0.002X+0.112, n=8, r2=0.766, F=19.620, P=0.004. Curved lines represent
95 % confidence intervals.

31

 

03 I I I I

Log Weighted Density

 

 

 

 

-01 l I I I
O 19 38 57 76 95

Large Cladoceran Density (it/L)

Figure 8 Linear regression plot of log weighted K. cochlearis typica density by large
cladoceran density for Day 5. K. cochlearis typica density weighted by Day 0 density.
Y=-0.002X+0.115, n=8, r2=0.887, F=47.212, P=0.0001. Curved lines represent

95 % confidence intervals.

32

 

0.3 I I I I

02 r 4

Log Weighted Density

 

 

 

 

-01 I I I I
O 19 38 57 76 95

Large Cladoceran Density (it/L)

Figure 9 Linear regression plot of log weighted K. cochlearis typica density by large
cladoceran density for Day 6. K. cochlean's typica density weighted by Day 0 density.
Y=-0.001X+0.107, n=8, r2=0.891, F=49.200, P=0.0001. Curved lines represent
95 % confidence intervals.

33

 

03 I I I I

Log Weighted Density

 

 

 

 

”0.1 l l l 4

O 19 38 57 76 95
Large Cladoceran Density (fr/L)

Figure 10 Linear regression plot of log weighted Polyarthra spp. density by large
cladoceran density for Day 5. Polyarrhra spp. density weighted by Day 0 density.
Y=-0.003X+0.253, n=8, r2=0.876, F=42.509, P=0.001. Curved lines represent
95% confidence intervals.

34

 

0.3 I I I I

O2 *- ._

 

Log Weighted Density

 

 

 

 

‘01 l l l l
O 19 38 57 76 95

Large Cladoceran Density (it/L)

Figure 11 Linear regression plot of log weighted Polyarthra spp. density by large
cladoceran density for Day 6. Polyarthra spp. density weighted by Day 0 density.
Y=-0.002X+0.156, n=8, r2=0.504, F=6.099, P=0.048. Curved lines represent
95 % confidence intervals.

population rate of change.

On both Days 5 and 6 (Figures 6—9), Anuraeopsis spp. and K. cochlean's
typica show clear trends of declining density with increased large cladoceran density.
These trends are all significant at the p<0.05 level. Polyanhra spp. densities show a
significant decline on Day 5 (Figure 10) (P<0.05), but while the trend seems to be
there on Day 6 (Figure 11), it is not significant.

Large cladocerans also suppressed small cladocerans. Figure 12 shows a
significant (P <0.05) negative trend for a linear regression of Ceriodaphnia reticulata
density by large cladoceran density. Bosmina longirosm's density also shows a
significant (P<0.05) negative trend by large cladoceran density (Figure 13). Large
Cladocerans cannot interfere with adult small cladocerans, thus it appears the

mechanism of suppression is exploitative competition.
Mortality Rates: observed and expected

To compare observed rotifer species dynamics to dynamics expected from
interference competition and exploitative competition, observed daily mortality rates
(M,) were calculated for Anuraeopsis spp. , K. cochlearis typica, Polyarthra spp. from

Day 0-5 data. These Mds were calculated as follows:

Np: 0(3) (for one day) (1)

35

Np: 0(32 (for 2 days) (2)

NF: 00535 (for 5 days) (3)
S=V(N,JN0) (4)
Md=(l -S) (5)

Where S= daily survivorship rate, Np=final density, and No=initial density.
Equation 1 shows the relationship between N, No, and S for one day. Equation 2
shows that NF for a second day is the number that survived from the first day
multiplied by the daily survivorship (S) again. Equation 3 extrapolates this
relationship to 5 days. Using equation 4, one can calculate the daily survivorship rate
(for 5 days) from initial and final densities. An Observed daily mortality rate (Md) for
this period can be calculated by subtracting the daily survivorship rate from 1
(Equation 5). Observed Mds are actually estimates of 1' since they include both
mortality and birth. Observed Mds are therefore conservative estimates of mortality
since any birth would cause an underestimation of mortality rate.

Observed Mas for Anuraeopsis spp. , K. cochlearis typica, and Polyarthra spp.
are plotted in Figure 14. Dependant paired t-test comparisons (Wilkinson 1990) of

mean changes in observed Mas from zero found K. cochlearis typica and Polyarthra

36

37

 

400 I I l I

ZOO

Ceriodaohnia Density (it/L)

lOO

 

 

 

O I I I I
O 19 38 57 76 95

Large Cladoceran Density (it/L)

 

figure 12 Linear regression plot of Ceriodaphnia reticulum density by large
cladoceran density (treatment). Y=-3.006X+267.083, n=8, r2=0.644, F =10.877,
P=0.016. Curved lines represent 95% confidence intervals.

38

 

200 I I I I

100

Bosmina Density (ti/L)

 

 

 

 

‘100 I I I I
O 19 38 57 76 95

Large Cladoceran Density (it/L)

Figure 13 Linear regression plot of Bosmina Iongirostris density by large cladoceran
density (treatment). Y=-0.895X+69.455, n=8, r2=0.672, F=12.274, P=0.013.
Curved lines represent 95 % confidence intervals.

spp. Mas to be significantly different from the Anuraeopsis spp. Md (P<0.05), but the
M,' of K. cochlearis typica and Polyarthra spp. were not found to be different (Table
4).

Mds expected solely by interference competition were calculated from a
magnified figure from Burns and Gilbert (1986b) (Their figure 6, not shown here).
Their figure, based on laboratory data, displays the relationship between daphnid body
length, daphnid density, and percent K. cochlearis tecta killed per day. Percent K.
cochlearis tecta killed per day values (my Mas) were estimated from the magnified
figure for each of my experimental bag treatments (large cladoceran densities).
Keratella cochlearis tecta is a spineless form of K. cochlearis typica which is a model
interference susceptible rotifer species. Calculated Mas values were modified for K.
cochlearis typica and Polyarthra spp. based on their susceptibilities to interference
competition relative to K. cochlean's tecta reported by Gilbert (1988b) (Gilbert’s table
3). Gilbert shows K. cochlearis typica (the spined morph present in my study) to be
about 65 % as susceptible to interference as K. cochlearis tecta, and Polyarthra remata
to be only 10% as susceptible as K. cochlearis tecta. The expected interference Mds
for K. cochlearis typica and Anuraeopsis spp. were therefore calculated as 65 % and
10% that of the K. cochlearis tecta values respectively. Anuraeopsis sp. was not
included in Gilbert’s study, but it is expected to be at least as susceptible to
interference competition as K. cochlearis tecta (1.]. Gilbert, pers. comm.) because of
its small size, weak lorica, and lack of a sophisticated escape response. K. cochlearis

tecta Mas were therefore used as Polyarthra spp. values.

39

40

Table 4 Paired t-test comparisons of mean change in observed daily mortality rate
from zero between three rotifer species. Tests if mortality rates are different. An

asterisk (*) denotes significance at the a=0.05 level.

 

 

 

 

Polyarthra

 

 

 

 

 

 

Comparison Mean Change N S.D. t p
Anuraeopsis vs 12.532 8 13.412 2.643 0.033 *
Keratella
Anuraeopsis vs 15.855 8 9.764 4.602 0.002 *
Polyarthra
Keratella vs 3.353 8 16.631 0.570 0.586

 

 

Figures 15-17 display Mds for observed data and Ms expected solely due to
interference competition for Anuraeopsis spp. , K. cochlearis typica, and Polyarthra
spp. respectively. Interference competition expected values did not meet the
assumptions of a dependent matched paired t-test (normality), so a Wilcoxon Signed
Ranks test (Table 5) was used to test for differences between observed and expected
Mds. (Wilkinson 1990). By both means, only Polyarthra spp. was significantly
different (P < 0.05). Therefore, K. cochlearis typica and Anuraeopsis spp. observed
Mds appear to be consistent with interference expected Mds, while Polyarthra spp.
Mas do not.

Gilbert ( 1985) and MacIsaac and Gilbert (1991a) conducted starvation
experiments with K. cochlearis typica and found M,I =0.23 ind ~d". There were no
other readily comparable Mds to be found in the literature, so a qualitative comparison
is made only with observed values for K. cochlearis typica. Starvation can be viewed
as the most extreme case of exploitative competition. This exploitative competition
expected M, is compared with Observed Mds in Figure 14 and appears to be
inconsistent with Observed Mds which are generally greater than the expected Mds.
This is especially true at higher large cladoceran densities where more intense
starvation-like exploitative effects should occur.

Egg ratios of Anuraeopsis spp. and K. cochlearis typica were examined by
linear regression across large cladoceran density by Day (Day 0,5,6 and pooled Day

5,6 data). Points for large cladoceran densities above 16 ind - L" were based upon too

41

42

Table 5 Wilcoxon Signed Ranks test comparing Observed vs. Expected daily
mortality rates within each species. An asterisk (*) denotes significance at the a=0.05
level.

 

 

 

 

 

 

 

 

 

Species Variable A Z p
Anuraeopsis spp. Observed 5 0. 845 0.398
Expected 3
Keratella cochlearis typica Observed 5 0.140 0. 889
Expected 3
Polyarthra spp. Observed 7 -2.240 0.025 *
Expected 1

 

 

43

 

 

 

 

120 r a I I
100
g
93 80
CO
CC
3:
E 60
5
E
g 4O
(0
D
............. Starve
20 -------- Poly
— — — Kcoc
O Anur

 

 

Large Cladoceran Density (it/L)

Figure 14 Observed daily mortality rates, Md, (% killed per day) for Anuraeopsis
spp. , K. cochlean's typica, and Polyarthra spp.. Rates based on Day 0—5. STARVE is
the line for expected daily mortaliy rates from extreme exploitative competition based
on Gilbert’s (1985) starvation experiment with K. cochlearis tecta.

 

 

 

 

 

 

120 T r I .
lOO -
.3
g 80 '7
(U
CC
2:
IE 50 —
5
E
g 40 a
(U
Q
20 ~
----- Expected
O u I I I I Observed
O 19 38 57 76 95

Large Cladoceran Density (fir/L)

Figure 15 Anuraeopsis spp. Observed and Expected daily mortality rates, Md, (%
killed per day). Expected rates are based on interference effects solely (Burns and
Gilbert 1986b). Observed rates calculated from Day 0—5.

45

 

120 I I f I

100 T T

80— _.

Daily Mortality Rate (‘70)

 

----- Expected
Observed

 

 

 

 

O .. I I I I
O 19 38 67 76 95

Large Cladoceran Density (it/L)

Figure 16 K. cochlearis typica Observed and Expected daily mortality rates, Md (%
killed per day). Expected rates are based on interference effects solely (Burns and
Gilbert 1986b). Observed rates calculated from Day 0-5.

46

 

 

 

120 f I I I
100
g
g 80
(U
CC
3‘
E GO
5
2
g 40
CO
D
20
————— Expected
Observed

 

 

 

Large Cladoceran Density (it/L)

Figure 17 Polyarthra spp. Observed and Expected daily mortality rates, M,I (% Killed
per day). Expected rates are based on interference effects solely (Burns and Gilbert
1986b). Observed rates calculated from Day 0—5.

few rotifer individuals counted (1-22 inds), so only data points for the 0-16 ind -L"
large cladoceran density range were used. All of the points were derived from at least
40 individual rotifers counted.

Data are plotted by day for K. cochlearis typica and Anuraeopsis spp. in
Figures 18-20 and 21-23 respectively, and show a marked decline in the absolute
value of egg ratios over time. Because Day 5 and 6 egg ratios for K. cochlearis
typica and Anuraeopsis spp. were very close to zero, matched paired t-test
comparisons were performed (Table 6) to determine if egg ratios are significantly
different from zero (ie. is there zero effective reproduction occurring). On Days 0 and
5 (Figures 18 & 19), K. cochlearis typica egg ratios were found to be significantly
different from zero (P<0.05), as was Anuraeopsis spp. for Day 0 (Figure 21).
Paired t-test comparisons were also conducted upon averaged, pooled egg ratio
values for Day 5 and 6 of each species in an effort to account for variability in egg
ratio value between days. Pooling produced significant results only in K. cochlearis
typica (Table 6). These results indicate that for Anuraeopsis spp. on Day 0 and for K.
cochlearis typica on Day 0, 5, and pooled Days 5&6, significant reproduction is
occurring.

Regressions on egg ratio data for K. cochlearis typica and Anuraeopsis spp. by
day across treatments (large cladoceran density) found none of the slopes to be
significantly different from zero (Figures 18-20, and 21-23 respectively), implying
that egg ratios did not differ significantly across the large cladoceran density gradient.

The highest large cladoceran density (16.4 inds - L") exhibited high leverage values,

47

48

Table 6 Paired t-test comparisons of mean change in egg ratio and zero from zero.
Tests if egg ratios are significantly different from zero. Day = 5&6 is the lumped
average of Day = 5 and Day = 6. An asterisk (*) denotes significance at the a=0.05

 

 

 

level.
Species egg ratio Day Mean N S.D. t p H
vs 0 comparison Change
Anuraeopsis spp. 0 -0.365 5 0.038 -21.270 0.0001
vs 0 5 -0.080 5 0.104 -1.715 0.161
6 —0.111 5 0.137 -1.811 0.144
5&6 -0.095 5 0.116 -1.830 0.141
I Keratella cochlearis 0 -0.287 5 0.041 -15.730 0.0001
typica 5 -0.032 5 0.024 -3.064 0.038
6 5 -2.368 0.077
& -2.925 0.043

 

 

 

 

 

 

 

 

fl

 

49

 

 

 

 

 

.9
E 0
Ct 02 — -
O)
O)
LlJ
OI 3
OO - _
-O.1 I I 4
-lO 0 IO 20 so

Large Cladoceran Density (it/L)

Figure 18 Plot of linear regression on Egg ratio for K. cochlearis typica for Day 0.
Y=0.001X+0.284, n=5, P=0.022, F=0.066, P=0.814. Curved lines represent

95 % confidence intervals.

50

 

0.5 I I I

OS ~ —
(12 - -

(JO E 0 —

Egg Ratio

 

 

 

 

“’01 L l 1
—IO 0 IO 20 SO

Large Cladoceran Density (It/L)

Figure 19 Plot of linear regression on egg ratio for K. cochlearis typica Day 5.
Y=-0.001X+0.036, n=5, r2=0.072, F=0.231, P=0.664. Curved lines represent
95 % confidence intervals.

51

 

 

 

0.5 I I I
O4 - ~
0.3 — _
.9
E3
0)
CD
LLl
OI - D _
00 N.
-01 /I—I\I\

 

—10 O 10 20 30
Large Cladoceran Density (it/L)

Figure 20 Plot of linear regression on Egg ratio for K. cochlearis typica for Day 6
Y=-0.002X+0.056, n=5, r2=0.358, F=1.672, P=0.287. Curved lines represent
95 % confidence intervals.

52

 

0.5 V

0.4" 00 T

 

()2 T

Egg Raho

0.0 E I

 

 

 

-01 I I I
-lO O 10 2O 30

Large Cladoceran Density (It/L)

Figure 21 Plot of linear regression on Egg ratio for Anuraeopsis spp. for Day 0.
Y=-000X+0.365, n=5, r2=0.OOO, F=0.001, P=0.981. Curved lines represent 95%

confidence intervals.

53

 

Egg Ratio

 

 

 

 

—O.l
—IO O 10 20 SO

Large Cladoceran Density (It/L)

Figure 22 Plot of linear regression on Egg ratio for Anuraeopsis spp. for Day 5 .
Y=-0.005X+0.116, n=5, r2=0.305, F=1.3l7, P=0.334. Curved lines represent

95 % confidence intervals.

54

 

0.5 I

OCB -

Egg Raho

 

 

 

 

-Ol
-lO O IO 20 SO

Large Cladoceran Density (it/L)

Figure 23 Plot of linear regression on Egg ratio for Anuraeopsis spp. for Day 6.
Y=-0.006X+0.156, n=5, r2=0.279, F=1.160, P=0.360. Curved lines represent

95 % confidence intervals.

signifying that it exerted particular influence over the regression. Regressions were
run without this point (not shown) leading to the same result, so the 16.4 inds -L"
points were left in the analysis and accepted as valid data.

Since the null hypothesis of slope=0 for the egg ratio regressions could not be
rejected at P<0.05 for any of the regressions, and the number of points in the
regressions were low along with very low Day 5&6 values and rzs , a power analysis
(Gerrodette 1987) was conducted on the Day 5 K. cochlear-is typica regression. This
day was the only single day occurrence of a non-rejected null hypothesis whose values
were significantly different from zero. The power analysis determined the power with
which one could detect the largest biologically interesting difference from the
observed regression line. The largest biologically interesting difference is represented
by the line which results in zero reproduction at the highest large cladoceran density
(16 ind -L"). This is interesting because it represents a significant negative trend Of
egg ratio with large cladoceran density. The power analysis therefore determines
whether one has the power to detect such a significant trend (”power” line). This
”power” line was calculated from equations in Gerrodette 1987 to have a slope equal
to -0.0247. The power to detect was set at 0.7. This power analysis is graphically
represented in Figure 24, and was meant to give some indication of the confidence
with which one can adopt the null hypothesis (A.J. Tessier pers comm.). The power
to detect such a difference was calculated to be 0.04. which means that there is a 96%
chance of committing a Type 11 error (incorrectly rejecting the null hypothesis (Gill

1978)).

55

Egg Ratio

020

0.15

0.10

0.05

OOO

-0.05

56

 

 

 

 

l I I

I‘\\\
\\‘ D
‘%\s
D \\
\\‘ D
\\\‘\\
r— o ‘m
/—J\ P

10 O IO 20

Large Cladoceran Density (tit/L)

BO

Power
Observed

 

Figure 24 Plot of K. cochlearis typica Day 5 power analysis on Egg ratio showing

Observed (Y=-0.001X+0.036) and the largest interesting detectable difference from
Observed (Power, Y=-0.002X+0.057). Power to detect (l-B)=0.035. Curved lines
represent 95 % confidence intervals.

DISCUSSION

A common problem in working with bag enclosures in the pelagic zone is that
bags often take on characteristics which are very different from the system being
modelled. These bag effects are mainly caused by interactions of the artificially
contained "limnetic" water with the sides of the bag. By creating substrates (bag
material) adjacent to bag water, periphyton growth ensues which can eventually
change the nature of the bag enclosure . Separation of bag water from lake water also
inhibits wind induced lake currents that enhance the suspension of particles.
Separation also precludes gaseous exchange and nutrient flow between bag water and
lake water (Bloesch et al. 1988). These bag effects are both a function of time and
lake trophic status. As time progresses, bag walls become progressively colonized,
causing interactions with bag water which are not typical of limnetic zones. Lake
trophic status can effect the rate of bag effect manifestations because systems with
more nutrients colonize bag walls faster. Thus, there is variation in the point in time
at which bag effects become problematic to limnetic studies.

In the present study, bag effects were manifested during the course of the
experiment. The drop in the absolute value of rotifer egg ratios (Figure 18-23) can
probably be attributed to settling of algae. This is inferred because all treatments
regardless of large cladoceran density exhibited such decreases. This hypothesis is
also somewhat consistent with comparisons of secchi depths between bag enclosure

and Wintergreen lake. Secchi depths for the bags were greater than the lake (~0.8-

57

1.6 M vs. 0.7 M). Differences in secchi depths between bags were probably due to
large cladoceran clearance of filamentous cyanobacteria from the water column.
Filamentous cyanobacteria is positively buoyant because it contains gas vacuoles
(Darley 1982), therefore filamentous cyanobacteria would not settle out in the bags as
other algal is presumed to have done. Bags were open at the top. Bag temperatures
with depth were not different from values in the lake.

Examination of accuracy in rotifer density counts via coefficients of variation
for between-subsamples and between-samples (Table 3) show that between-subsample
count coefficients appear to be somewhat lower than between-samples. These data
suggest that the subsampling method was adequate, though rotifers may have been
patchy in their distribution within the bag. Patchiness would be horizontal patchiness,
because the sampler took vertically integrated samples.

Unforeseen problems with establishing large cladoceran density treatments and
maintenance of the treatment integrity (bag loss etc.) resulted in lack of replication
and a reliance upon linear regression methods of analysis. Non-replicated large
cladoceran density treatments were successfully established (Table 2), and provided a
meaningful large cladoceran density gradient to address the experiment’s central
hypotheses. All treatments except for the control contained large cladocerans of more

than adequate size (1.71 i024 mm) to exhibit interference competition effects.

58

Rotifer suppression

The results of the log weighted density regressions clearly show that large
cladocerans have a suppression effect upon all three species of rotifer examined
(Anuraeopsis spp. , K. cochlearis typica, Polyarthra spp.)(Figure 6—11).

The response of small cladocerans also suggests a suppression effect by large
cladocerans (Figures 12-13). If suppression is due to large cladoceran density as it
appears, these data suggest that exploitative competition is high. Interference
competition, via mechanical interference, is not possible in this situation because
small cladocerans are not small enough to be mechanically interfered with by large
cladocerans. Goulden et al. (1982) found that under low food conditions, B.
longirostris populations declined overtime in a manner consistent with Figure 13. The
interpretation of large cladoceran induced exploitative competition on small
cladocerans is potentially problematic because initial small cladoceran densities are
not known. Stocking variation was found in initial rotifer densities, requiring
weighting of their densities before analysis. However, it is still unlikely that stocking
variation can totally account for the negative trend in small cladoceran density with
increasing large cladoceran density. It is more likely that exploitative competition
suppressed small cladoceran populations.

It is possible that the small cladocerans could have been more susceptible to

the effects of exploitative competition with large cladocerans than rotifers were.

59

MacIsaac and Gilbert (1989) found in laboratory studies that K. cochlearis tecta was
able to competitively exclude the small cladoceran D. ambigua, and persisted for 8
weeks in moderate food level cultures with Ceriodaphnia dubia and B longirostris.
This suggests the possibility that some rotifers may be able to outcompete small
cladocerans for limiting resources.

When examining suppression (via daily mortality rates, M.) of individual
rotifer groups in comparison to each other, we find that Anuraeopsis spp. is
significantly more suppressed by large cladoceran density than K. cochlearis typica
and Polyarthra spp. (Table 4, Figure 14). These data are consistent with the
hypothesis of differential suppression via interference competition. If one assumes that
species which are more resistant to interference competition are also less resistant to
exploitative competition because of costs associated with diverting resources into
interference resistance structures (escape response apparatus, spines etc.), then the
data are not necessarily consistent with the hypothesis of differential suppression via
exploitative competition. If differential suppression via exploitative competition were
occurring, one would expect that Anuraeopsis spp. would be the least suppressed.

Comparisons of observed vs. interference competition expected Mds for each
of the three rotifer groups show that only Polyarthra spp. deviates significantly from
its expected Mas (Table 5, Figures 15—17). Polyarthra spp. exhibited a much higher
M, than would be expected. This difference may possibly be due to two factors. First,
expected Mas may have been overestimated. Estimates were based on laboratory

studies by Gilbert (1988b). In theses studies Polyarrhra spp. was found to be about 10

60

times as resistant as the highly susceptible model K. cochlearis tecta. These
laboratory results may overestimate resistance to interference competition because
they do not accurately model natural conditions. Under such laboratory
conditions,large cladocerans and rotifers were cultured in 50-250 m1 flasks, with
densities of about 1000 inds - L" and 25-50 inds - L" respectively. large cladoceran
vertical migration is prohibited under these conditions, resulting in elevated
interference competition exposure of rotifers. While Anuraeopsis spp. and K.
cochlearis M, estimates are also prone to this overestimation, the magnitude of
overestimation for Polyarthra spp. may be much greater because it’s interference
competition defense mechanism is an escape response while Anuraeopsis spp. and K.
cochlearis typica’s interference competition defense mechanisms are morphological.
Thus the Polyarrhra spp. interference competition defense mechanism is behaviorally
based and there may be some heightened escape response behavior associated with
artificial conditions which may not appear in nature. Anuraeopsis spp. and K.
cochlearis typica do not have significant behaviorally based escape responses, and
they are likely to be repetitively drawn into the branchial cavity, potentially suffering
mechanical damage both under laboratory and natural conditions.

Second, if there is a sufficient metabolic cost associated with elaborate escape
response apparatus, then Polyarthra spp. are likely to be more prone to limiting
resource effects. By Day 5 of the experiment, food levels (as judged by Anuraeopsis
spp. and K. cochlearis typica egg ratios; recall that egg ratios couldn’t be calculated

for Polyarthra spp.) were at limiting levels. Exploitative competition effects of large

61

cladoceran density were therefore exaggerated by algal settling and were likely to
afflict Polyarthra spp. the most.

In comparing observed vs. exploitative competition expected Mds (Figure 14) it
appears that there is more going on in the system than just exploitative competition.
Gilbert (1985) conducted starvation experiments upon K. cochlearis typica and found
an Md= 0.23. Starvation can be viewed as the most extreme case of exploitative
competition. Figure 14 shows a clear trend of increased M, with increasing large
cladoceran density for all three species groups, especially K. cochlearis typica.

Examination of egg ratio data yield somewhat less clear results. None of the
regressions for Anuraeopsis spp. and K. cochlearis typica show slopes significantly
different from zero (Figures 18-23), implying that egg ratios do not change across
large cladoceran density. If this were the case, it would be strong evidence that
exploitative competition is not occurring because all treatments would have similar
food availabilities (inferred from egg ratios). But the data are not strong enough to
make this argument convincing. The regressions are only based Off of 5 points, but
they do cover the range at which interference competition effects are expected to
become important (Burns and Gilbert 1986b). Day 0 egg ratios for Anuraeopsis spp.
and K. cochlearis typica are relatively high, but values for Days 5 and 6 are very low
probably due to algal settling. Only values for K. cochlean's typica on Day 5 and the
pooled data for Day 5 and 6 are significantly different from zero reproduction. A
power analysis of the Day 5 K. cochlearis typica data showed very low power for

detecting meaningful trends (Figure 24). At first blush it appears that egg ratios

62

across large cladoceran density treatments do not differ, but it seems premature to
suggest that these particular data show little exploitative competition occurring.

In summary, it is important to note three main mechanisms controlling
dynamics in this experiment. The first two are related to food limitation. Rotifers and
small cladocerans both show evidence of suffering food limitation effects caused by
algal settling and to a large degree exploitative competition from large cladocerans.
There is also evidence in differential rotifer mortality data, observed vs. expected
mortality data, and egg ratio data (in a much more confounded sense), that is
consistent with the hypothesis of interference competition occurring.

Thus as in most natural large cladoceran/rotifer systems, a picture develops Of
co-occurring interference competition and exploitative competition suggesting different
effects related to their magnitudes. Results of exploitative competition and
interference competition in the long term may be expected to depend greatly upon
initial environmental conditions (resource availability, etc.) and subsequent
successional events in the zooplankton community.

While exploitative competition effects may be exaggerated by herbivorous
members of the zooplankton community aside from large cladocerans, interference
effects should not be. Zooplankters such as omnivorous copepods do not filter feed in
the manner that large cladocerans do. Copepods are much more selective feeders,
selecting individual particles from appendage induced water currents (Koehl and
Strickler 1981). Copepods do not have branchial chambers as do cladocerans, so

rotifers that may become entrained into copepod feeding currents are not exposed to

63

feeding appendage movements for any appreciable time. Recall that large cladoceran
induced interference effects mainly depend upon exposure time to repeated feeding
appendage movements (residence time in the branchial chamber). Copepods are likely
to affect rotifer communities only by exploitative means and to some extent by
predation upon soft-bodied (illoricate) rotifers (Gilbert and Williamson 197 8,
Williamson 1983, Williamson and Butler 1986).

In the long term, conditions potentially governed by interference competition
(ie. with moderately high large cladoceran densities and high food availability) do not
appear to remain for long without additional mechanisms that maintain high food
availability. Without such mechanisms, large cladoceran populations would be
expected to increase until food resources are driven to limiting levels for large
cladocerans, and large cladoceran populations would decrease. In the time period
before food becomes limiting, rotifer reproduction should remain high with rotifer
suppression caused by interference competition effects. At this point rotifer species
relatively unsusceptible to interference competition should be selected for. Once
resources have become limiting, the competitive emphasis should switch to
exploitative competition, negatively impacting rotifer reproduction. As large
cladoceran populations decline with limiting food levels, rotifer species relatively
unsusceptible to exploitative competition should be selected for. Rotifer populations
may increase in size, if resource pulses occur (caused by nutrient pulses such as
runoff etc.), by virtue of their fast reproductive response. Interference competition

can potentially act as a gatekeeper, controlling the magnitude of these rotifer pulses.

64

This successional scenario could be changed, however, if large cladocerans are
prevented from driving resources to limiting levels. Mechanisms that interfere with
large cladoceran feeding efficiency (ie. suspended clays, high cyanobacterial
concentrations) may cause this change (Gilbert 1990, Gilbert and Durand 1990, Kirk
and Gilbert 1990). It is possible that these mechanisms may interfere with large
cladoceran feeding efficiency to such an extent that clearance rates are also reduced.
This would result in reduced interference competition.

Overall, in natural systems the interplay between exploitative competition and
interference competition appears to be potentially important in the successional
dynamics of dimictic temperate lakes. Rotifers may dominate in the spring (especially
in mesotrophic and eutrophic systems) by rapid reproductive responses to algal
blooms resulting from nutrients brought about by the spring tum-over. As large
cladoceran populations grow to a sufficient size in the late spring/early summer, they
competitively exclude rotifers by both exploitative and interference competition
depending on the level of food resources available. Large cladocerans dominate the
system throughout the summer once lake stratification occurs until the fall tum-over,
with possible brief outbreaks of rotifer abundance. In this successional scenario,
exploitative competition is expected to be the dominant form of competition with
interference competition playing a reinforcing role.

The dynamics of this seasonal succession should vary in lakes depending upon
large cladoceran densities. In lakes where large cladoceran densities are kept

relatively low due to moderate size-selective predation by planktivorous fish, rotifers

65

would be expected to persist until resources became overgrazed or the phytoplankton
community shifted to inedible types. The magnitude of population increases and
decreases in the successional scenario will also depend greatly upon the trophic status
of the lake. Oligotrophic lakes will not support large spring phytoplankton blooms like
eutrophic lakes will, because they are not as rich in nutrients.

The above discussion deals mainly with lakes, but interference and exploitative
competition should also occur in ponds. Their effects may not be as important as pond
environmental conditions in structuring the zooplankton community over time.

Ponds are not as well buffered as lakes are to environmental changes, thus
temperature, oxygen content, nutrient availability and algal population dynamics may
change more rapidly. These changes may affect cladoceran and rotifer population
growth much more than competitive effects. Interference and exploitative effects
should still be important when large cladocerans occur in adequate densities. Ponds
may be expected to be much more erratic in their cladoceran/rotifer dynamics due to

changes in pond environmental conditions.

Chapter III: Long Term Experiment

The purpose of this experiment was to assess the long term effects of
interference and exploitative competition upon a natural, diverse rotifer assemblage,
imposed by a nascent yet increasing population of large cladocerans. This large
cladoceran population introduced at low density was meant to model either the
seasonal occurrence of large cladoceran population growth in undisturbed systems or
large cladoceran introductions into disturbed systems which support large cladoceran
population growth but are dominated by rotifers and small cladocerans. The latter
scenario seems to be happening in Wintergreen Lake in the spring and summer of
1992 (G.G. Mittelbach, pers. comm.). Large cladoceran survival in the lake appears
to be caused by a marked decrease in size-selective predation pressure by golden
shiners (Notemigonous crysoleucas). Reduction in predation pressure results from the
successful reintroduction of the piscivorous largemouth bass (Micropterus salmoides).
In particular, this experiment was designed to test the hypothesis that in the long
term, rotifer species less susceptible to interference or exploitation by large
cladocerans would persist through time, possibly increasing depending upon the
relative magnitudes of interference and exploitation competition.

The long term experiment was conceived as a 2 X 2 replicate bag enclosure
design, with treatments of large cladoceran additions (Bags #2 & #4) and no large

cladoceran controls (bags #1 & #3). This experiment differed significantly from the

67

short term experiment (Chapter H) in that it 1) was run considerably longer (24 Days
vs. 6 Days) to allow for population growth of suppressed and/or competitively
released rotifer species, 2) was stocked with low large cladoceran (D. pulicaria)
densities allowing for recruitment of large cladoceran neonates into the size threshold
for interference (1.2 mm), simulating a natural introduction, and 3) employed larger
bag enclosures to buffer the experiment from confounding bag effects (Bloesch et a1.
1988).
METHODS

On 8/5/91, 4 enclosure bags holding a volume of 1963 L (1 M in diameter,
2.5 M deep) each were introduced into Wintergreen Lake. Bags were made of 4 mil
polyethylene and were heat-sealed, cable-tied, and bricked as in the short term
experiment. Bags were open at the top, and suspended from wooden frames fitted
with a styrofoam collar. Over the top of each bag frame I fitted a wooden frame
holding a 2 ply panel of clear 4 mil polyethylene sheeting that allowed light
transmission. Covers were elevated 3 cm above the bag frames to allow for adequate
ventilation of the bags. Covers were used to prevent introduction of fecal matter
(nutrient input) from waterfowl using the enclosures as roosting sites. Bags were
tethered together and moored as in the short term experiment (Figure 5).

On 8/5/91 bags were filled by hand as in the short term experiment. A 5th bag
(no large cladoceran control) was introduced on 8/9/91 because bag #1 was losing
volume. This demonic intrusion was exorcised with the introduction of bag #5, as bag

volume soon stabilized. The volume loss of bag #1 was not so great as to preclude its

68

usefulness, and it was monitored in the rest of the experiment as was bag #5. Bag #1
data was used in the analysis of this experiment while bag #5 data was not. Bag #5
was excluded because it was not started at the same time as the others, it was 3 days
younger at any given time. This bag age difference would have caused unnecessary
interpretation of the affect of differential bag age. Bags #2 and #4 were stocked at a

density of 0.5 inds-L" with D. pulicaria from Gull Lake, MI. as in the short term

experiment.

Sample Collection and Preservation

Rotifers and cladocerans were sampled simultaneously every three days (except
for one instance where the period was four days) using a 9 L Van Dorn bottle
sampler. Before sampling, a secchi disk was lowered to the bottom of each bag and
quickly drawn to the surface. This process was repeated 4-5 times to mix the bag’s
content thoroughly, thus disrupting any patchiness in the zooplankton distribution
(Geedey 1990). It was hoped that this procedure would also serve to resuspend settled
algae. Observations of before and after mixing seemed to support the resuspension
hypothesis because bag water was noticeably more turbid after mixing.

A sample was then taken at a depth of 1 M, and was filtered through a 40 pm
mesh plankton funnel (Likens and Gilbert 1970). Plankton were transferred into
labelled 150 ml QuorpakTM bottles, stored on ice, transported to the laboratory, and

preserved with 5 % cold sugar formalin. Two samples were taken per bag, mixing the

69

bag as above before each sample. Filtered water was returned to the enclosure.

Algal samples, 02 profiles, and temperature profrles were taken once a week
as in the short term experiment. Temperature and 02 measurements were taken on
only 2 of the bags each week (one test, one control).

On 8/17/91 (Day 11 of the experiment) a storm occurred on Wintergreen Lake
which removed enclosure covers for a period of 2.5 days. As a result of this event,
bags were exposed to a prodigious input of waterfowl feces since many waterfowl
perched on the bag floats. This was confirmed visually. When the enclosure bags
were mixed for sampling, suspended particulate fecal matter caused the plankton
funnel to become completely clogged rendering the sampling procedure infeasible.
Fecal inputs may also have caused decreased light penetration and increased nutrients.
To enable continued sampling thereafter, bags were not mixed and sampling was
conducted at night between 2200-2400 hrs. Night sampling was initiated in order to
capture vertically migrating large cladocerans which would have been at the bag
bottom during the day. Samples were otherwise collected and processed in a similar

manner as pre-storm samples were.

Sample Processing

Rotifers and cladocerans (large and small) were processed from the same
sample, but were processed and counted differently. Rotifer samples for each

sampling date were processed. They were concentrated using a 45 um screen, brought

70

to a standard volume of 100 mls, randomly mixed, and 1 ml subsamples were taken
using a wide-bore pipet. Subsamples were placed in a Sedgewick—Rafter cell for
enumeration. Five (5) random lengthwise transects across each cell were counted for
all rotifer species at 100x. Replicate subsamples were taken until at least 25
individuals from each of the three major species groups (Anuraeopsis spp. , K.
cochlearis typica, and Polyan‘hra spp.) were counted. At least two subsamples were
always counted, however samples beyond day 10 were not counted (except for day
24, see below). The number of eggs carried by individuals was also recorded.

Large cladocerans, small cladocerans, and copepods were processed and
enumerated as in the short term experiment, but only weekly sampling dates were

processed. Algal samples for this experiment were never processed.

RESULTS AND DISCUSSION

Most of the data for this experiment were not vigorously analyzed because the
instituted treatments (large cladoceran density) never attained a magnitude desired to
produce strong interference competition effects and true replication was not achieved.
During the course of the experiment some Ceriodaphnia reticulara individuals grew
above the 1.2 mm size threshold. C. reticulata was a common small cladoceran in
both this experiment and the short term experiment, but never grew above 1.2 mm in
the short term experiment. Bag #2 peaked at a total large cladoceran (D. pulicarla and

C. reticulata) density of 3.28 inds - L"(at day 10) which declined to less than 1

71

inds-L" by day 24. All other bags peaked at less than 1.5 inds -L" during the first 10
days (Table 7). It can be argued that C. reticulum individuals larger than 1.2 mm do
not have a carapace shape conducive to permitting effective interference of rotifers.
Even if one were to include them though, at such low large cladoceran densities
interference effects are expected to be very small (Burns and Gilbert 1986b). In
addition, if large C. rericulata can exert interference effects, the presence of large C.
reticulum ( > 1.22 mm) in the control bags (#1 & #3) may also represent a
confounding element potentially causing interference effects in the control bags. On
Day 0 of the long term experiment, no large C. reticulara individuals were found
(Table 7). Large C. reticulata did not appear in the bags until sometime after Day 3.
By Day 10 large C. reticulata attained densities in bag #1 greater than the densities of
large D. pulicaria. Thus it seems that the experiment shows little promise of
adequately addressing the hypotheses tested.

Upon examination of the rotifer samples, I found that after Day 10, samples
contained increasingly higher numbers of littoral rotifer species (Lecane sp. ,
Lepadella sp. , Monortyla spp. , etc.) , that are not truly planktonic (Stemberger 1979).
These species did not represent normal constituents of the limnetic environment of
Wintergreen Lake (Fradkin pers. obs.). Luxuriant periphyton growth was also
observed on the sides of bags (up to 10 cm in length). Littoral rotifers were most
likely associated with the littoral-like properties of the periphyton. These observations
led me to believe that bag effects had sufficiently altered the nature of the bags,

rendering them inappropriate for use in this study. With the exception of Day 24

72

73

Table 7 Long term experiment Large-bodied Cladoceran Densities by Day and
Treatment number.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

" Treatment # Day D. pulicaria C. reticulata Total per
per liter per liter liter

II 1 3 0.00 0.00 0.00

I} 3 3 0.00 0.00 0.00
2 3 0.17 0.00 0.17
4 3 0.11 0.06 0.17
1 10 0.00 1.33 1.33
3 10 0.00 0.44 0.44 '1
2 10 2.89 0.39 3.28
4 10 1.17 0.33 1.50
l 24 0.00 0.94 0.94
3 24 0.00 0.67 0.67
2 24 0.00 0.89 0.89

= 4 24 1.67 0.61 2.28

 

 

 

 

 

 

samples which were counted to confirm the ”littoralization" trend, quantitative counts
of samples past Day 10 were not conducted. Examination of the data from Day 24
provides evidence that the bags do not reflect desirable limnetic conditions. Almost all
planktonic rotifers (especially Anuraeopsis spp. , K. cochlearis typica, and Polyarthra
spp.) were virtually extinct (Figures 25-28). By day 24, Lepadella sp. were the
dominant rotifers in most bags.

Physical data for this experiment showed that conditions in the bags did not
differ much from lake conditions. 02 levels at all depths in the bags were above 5
mg - L" during the entire 24 Days. Temperatures from top to bottom varied from
~ 24-29°C. None of these data are expected to have contributed to the failure of
establishing desired test treatments, though they could have contributed to the
enhancement of bag effects.

In its proposed goal, this experiment was unsuccessful, yet it does yield
possible insights about food limitation in the short term experiment. Examination of
K. cochlearis typica’s egg ratio data (Figure 28) for all treatments show K.
cochlearis typica’s maintenance of relatively high egg ratios through Day 10. Bags
were mixed for sampling every three days tending to resuspend settled algae. Egg
ratio data are consistent with the idea that bag mixing can effectively counteract the
negative effects of presumed algal sinkage on rotifer reproductive rates as
demonstrated by the short term experiment.

Anuraeopsis spp. and K. cochlearis typica populations decrease in density over

time (Figures 25 & 26), while Polyarthra spp. populations experience noticeable

74

75

DBagéi
IBagS
BBagZ
DBagI

 

l

 

   

  

‘N‘N‘N‘N‘N‘\\\\\‘\\\\\\\\\\\\\\\\\N ,

are»orororororowororowe»overflow.»overcrorowowororowere»

 

 

 

onononononononononoue

 

 

 

V“\\‘\\\\\\\\\.‘\‘\\‘\‘\\\\‘\\\\‘\‘\\‘\‘a

nonononononenouonononononononouononenonOHOHOHOHOMOHOHOHOMOMOMOHonenononouononononouonou.

 

 

 

  
 
   

N‘“\\‘\‘\\‘\‘\“\‘\\‘\

enouououououououououououououonouououououonououo

 

 

 

1500

1000
500 h

3:: :88 ESE

Day

Figure 25 Long term experiment Anuraeopsis spp. density (#/ L) for each Bag by day.

76

 

 

 

 
  

  

 

 

   
 
 

 

 

 

4 3 2 l

g g g g

a a a a

B B B B

D s E D
_ a

l .S\\\‘\\\\‘\‘\\‘\‘\\\\\\\‘\‘ ,

renew.“onenononononoufiouonflouo“ononononono”Onenononononon.

V“\“\\\“\\\\‘\\\\‘\‘\\\.\\‘\\\.\\‘\\\\\\\\\

T nouououououououoHonououououououomononououououououono”.Honououououououououououo ,
H —

O O O O
O O O
5 O 5

4| 4|

3).: :88 550m

24

10

Day

Figure 26 Long term experiment K. cochlearis typica density (#l L) for each Bag by

Day.

77

DBag4
IBagS

EBagZ
DBagl

 

r

 

.

  
  

<\‘\\\\\‘\\\\\\N\\\\\\\

OOOOOOOOOO

 

 

 

N\\‘\\\\\\\\\‘\‘\\\\\\\‘\\
WenonowonOMOMQMOMOMQMOMQMOMQuench.

   
  

 

 

  

SN\\‘\\\\\\‘

"one”..uononouonononoMonomer:

     

 

 

4000

3000

2000

3:: £88 stem

1000

Day

Figure 27 Long term experiment Polyarthra spp. densities (#lL) for each Bag by day.

78

Bag 8
Q Bag 2
El Bag 1

IBag4
S

 

 

‘\\\\\\.\.\.\.\.\\.\\.\.\\.\.\.\\\\.\.\.\.\\.\\.\\.\\\.\N
oOOOOOOOOOOOODOMOtOO0000000505OD000000000.000000099000000000000.

 

  

 

 

“OMONOMOMOMONOMOMOMOMonowonononononowoNOMONONOMOMOMOMOMO

 

 

 

   

   
 

ouonououououououonououououououo.,

 

 

 

0.4

0.3

O
Oi

0:9 mam cam}

OO

Day

Figure 28 Long term experiment K. cochlearis typica mean egg ratios for each Bag

by Day.

growth. K. cochlearis typica egg ratios appear to remain somewhat constant, if not
increasing, over the 10 day period (Figure 28), at the same time K. cochlearis typica
densities decrease. This trend is consistent with expected interference effects. It is
also possible that other interactions occurred in the bags, perhaps interspecific
competition between rotifers based on differential utilization of resources. Bogdan and
Gilbert (1982) show that certain rotifers are generalists in feeding, while others such

as Polyarthra spp. specialize on certain food types.

79

Chapter IV Conclusions and Recommendations

Conclusions

Overall, both experiments (short and long term) were beset with obstacles and
vagaries that could possibly have been averted and/or controlled to some degree by
improved experimental design and execution. These difficulties fall mainly into three
categories; 1) difficulties associated with bag effects, 2) difficulties in establishing
satisfactory treatment effects, and 3) obstacles associated with sampling type, method,
and frequency.

In the short term experiment, bag effects were manifested by algal settling
which exaggerated food limitation. Small bag diameter size (20 cm) made use of
standard sampling devices difficult (Van Dorn bottles etc), and inhibited manual
mixing of bags. In the long term experiment, bag effects were manifested by dense
periphyton growth and a shift in the rotifer community towards predominantly littoral
species. This "littoralization" effect made the bags no longer useful as models of
planktonic (limnetic) systems.

In the short term experiment, true replication was not achieved (Table 2),
leading to the abandonment of an Anova analysis and the adoption of regression
analysis. Large cladoceran stocking was not effective, potentially caused by problems

associated with collection and transport mortality of large cladocerans, ineffective

80

stocking methods, and large cladoceran intolerance to Wintergreen Lake conditions,
possibly high cyanobacterial concentrations (Gilbert 1990, Gilbert and Durand 1990).
Large cladocerans (D. pulicaria) were stocked from Oligotrophic Gull Lake, which
doesn’t experience high cyanobacterial blooms. Conversely, Wintergreen Lake
experiences very high concentrations of cyanobacteria throughout the summer, and
cyanobacteria are present throughout the year (Hall and Ehlinger 1989, Fradkin pers.
obs.).

In the long term experiment, Ceriodaphnia reticulata grew above the 1.2 mm
size threshold for interference, potentially confounding control bags with test bags that
held both large C. reticulata and the stocked large D. pulicaria. In addition, D.
pulicaria populations did not increase in size in the way expected in order to establish
desired treatment effects (Table 7). D. pulicaria peaked at around 3 inds - L" and was
extinct by Day 24 in its largest treatment. Consideration of the large cladoceran
density data caused me to conclude that desired treatment effects were not achieved.

Data from both short term and long term experiments also suffered from
various sampling deficiencies. Egg ratio data from both experiments were used as an
indicator of food availability, which was used to infer the magnitude of exploitative
competition effects. In the short term experiment these data were not inconsistent with
the hypothesis of large cladoceran induced interference effects, but exhibited such low
power as to view their interpretation with severe skepticism. Exploitative effects were
exaggerated by and confounded with algal sinkage effects on rotifer reproduction. Egg

ratio data were the only major indicator of food availability because algal samples

81

were not processed. In addition, high coefficients of variation for sampling and
subsampling procedures (Table 3) suggest that these procedures were not wholly
adequate and should be improved upon.

Sampling frequency for rotifers (Days 0,5 & 6) and cladocerans (Day 6) in the
short term experiment required the assumption of constant daily mortality rates (M)
be made in M, calculations. It also required that final large cladoceran density be used
as the constant effective density throughout the experiment. While these assumptions
seem reasonable, more frequent sampling would have increased confidence in these
assumptions. Short term experiment coefficients of variance between samples were
relatively high, suggesting that the vertically integrated samples were not wholly
accurate in sampling possible horizontally patchy rotifers. As previously stated, small
bag diameter limited the type of sampler available for use.

Sampling frequency for the long term experiment (~ 3 days) was probably
adequate, but the shift in sampling strategy from day/bag-mixing to night/non-mixing
would have made comparison of data derived from these schemes challenging even if
the aforementioned bag effects had not appeared. As in the short term experiment,
algal samples were not processed making the egg ratios the only indicators of large

cladoceran induced exploitative effects.

82

Recommendations

If the above experiments were to be conducted again, several significant steps
could be taken to help resolve some of the above mentioned problems. For the short
term experiments, bag effects could be temporarily brought under control and more
meaningful data gathered by; 1) increasing the bag size, 2) increasing the rotifer and
cladoceran sampling frequency, 3) regularly mixing the bags, 4) changing sampling
device and sample size, 5) improving the efficiency of large cladoceran stocking or
possibly switching to a more robust species/clone from a eutrophic lake, 6) analyzing
algal data, and 7) monitoring zooplankton dynamics in Wintergreen Lake concurrently
with the experiment.

By increasing the bag size, the problem of bag ”littoralization" (bag material
becoming substrate for periphyton causing the bag to deviate from limnetic
conditions) should become diluted as the bag surface to volume ratio increases.
Increased bag size also allows for easier manual bag mixing (up to a point) and use of
standard sampling devices. By using a device such a Van Dom bottle for sampling
rotifers from mixed bags, small between sample coefficients of variation should
result. An increased sample volume would aslo serve to decrease coefficients Of
variation. Increased sampling frequency would also be enabled by increased bag size
because samples would represent a smaller percentage of the total bag volume
harvested. Thus, increased sampling frequency should be expected to give a clearer

picture of zooplankton dynamics over the 6 day time period.

83

More efficient large cladoceran stocking would enable treatment replication
and use of more powerful statistical analyses (ie. repeated measures Anovas).
Efficient stocking may be accomplished by attempting to reduce stress on the large
cladocerans during the process of collection, processing, and transport to stocking
site. A shorter time interval between collection and stocking would help. It would also
be prudent to collect large cladocerans from sites with a more similar lake trophic
status to Wintergreen Lake, in order to minimize possible negative cyanobacterial
concentration effects. Large batch cultures of individuals could possibly be maintained
under Wintergreen Lake conditions to acclimate large cladocerans, thought this would
be labor intensive and infeasible if bag sizes are too large. The number of large
cladocerans needed for stocking would be too great to culture feasibly. Altemately, a
large cladoceran species other than D. pulicaria could be used if it showed better
promise of survivorship. Hall and Ehlinger (1989) demonstrated that D. galeata
mendotae (the other dominant large cladoceran in pre-winterkill Wintergreen Lake)
can thrive under Wintergreen Lake conditions. This species was used in a trial run of
the short term experiment with poor results. These results can possibly be attributed
to its affinity for entrapment in the surface film (A.J. Tessier, pers. comm.). In
addition, these individuals were collected from Oligotrophic Lawrence Lake, MI and
may have had difficulties in acclimating to Wintergreen Iake’s high cyanobacterial
concentrations.

An analysis of algal samples would give added evidence of food availability,

potentially bolstering egg ratio data. Egg ratio data itself would become more

84

powerful as a result of bag mixing. Increasing the numbers of individuals counted
would also increase the reliability of egg ratio data. Quantitative algal cell counts
would provide data on the relative ratio of edible to inedible food resources, while
chlorophyll counts would provide total algal biomass data (Wetzel and Likens 1991).
These algal data use in conjunction with egg ratio data should give a clearer picture of
exploitative competition effects in the bag systems.

Monitoring Wintergreen Lake zmplankton dynamics concurrently with the
experiment would give an indication to what extent bag dynamics of control
treatments reflect the nature of lake dynamics. Such monitoring may also be used to
help explain unexpected bag dynamics caused by environmental variables etc.

Modification of the long term experimental design to make it successful would
be significantly more challenging than modification of the short term experiment.
Short of moving the experiment to a less eutrophic system, I believe bag effects
would most likely alter the character of the bags well before meaningful long term
results could be obtained. Periphyton build-up on the bag sides could be removed
manually, but this procedure could prove logistically difficult and could also
potentially introduce large amounts of particulate matter into the water column.

Movement of the long term experiment to large ( ~ 10,000 L) mesotrophic
tanks, with natural stocked rotifer, large cladoceran, and phytoplankton communities
offers reasonable prospects for success. These tanks would need to be mixed to inhibit
algal settling, and might also have to be sampled in several locations in order to

ensure representative samples if the tanks could not be easily be mixed

85

homogeneously for zooplankton. Tanks could be set up with replicate tests and
controls, with care given to assuring stocked large cladoceran densities are accurate as
outlined above. An undertaking of this magnitude was unfortunately outside of the

realm of this study.

86

BIBLIOGRAPHY

Bibliography

Adalsteinsson, H. 1979. Zooplankton and its relation to available food in Lake
Myvatn. Oikos 32:162-194.

Bloesch, J., P. Bossard, H. Biihrer, H. R. Biirgi, and U. Uehlinger. 1988. Can
results from limnocorral experiments be transferred to in situ conditions?
Hydrobiologia 159:297-308.

Bogdan, K. G., and J. J. Gilbert. 1982. Seasonal patterns of feeding by natural
populations of Keratella, Polyarthra, and Bosmina: Clearance rates,
selectivities, and contributions to community grazing. Limnology and
Oceanography 27(5):918-934.

Bogdan, K. G., J. J. Gilbert, and P. L. Starkweather. 1980. In situ clearance rates
of planktonic rotifers. Hydrobiologia 73:73-77.

Burns, C. W., and J. J. Gilbert. 1986a. Direct observations of the mechanisms of
interference between Daphnia and Keratella cochlearis. Limnology and
Oceanography 31(4):859-866.

---. 1986b. Effects of daphnid size and density on interference between Daphnia and
Keratella cochlearis. Limnology and Oceanography 31(4):848—858.

Byron, E. R., P. E. Sawyer, and C. R. Goldman. 1986. The recurrence of Daphnia
rosea in Lake Tahoe: analysis of a population pulse. Journal of Plankton
Research 8:771-783.

Carpenter, S. R., J. F. Kitchell, and J. R. Hodgson. 1985. Cascading trophic
interactions and lake productivity. Bioscience 35:634-639.

Dabom, G. R., J. A. Hayward, and T. E. Quinney. 1978. Studies on Daphnia
pulex Leydig in sewage oxidation ponds. Canadian Journal of Fisheries and
Aquatic Sciences 56: 1392-1401.

Darley, W. M. 1982. Algal biology: a physiological approach, 1st Edition. Volume
9. Blackwell Scientific Publications, Oxford.

DeMott, W. R. 1989. The role of competition in zooplankton succession. Pages
195-252 in U. Sommer, editor. Plankton Ecology. Springer-Verlag, Berlin.

87

Edmondson, W. T. 1965. Reproductive rate of planktonic rotifers as related to
food and termperature in nature. Ecological Monographs 35:61-11].

Gwdey, C. K. 1990. The seasonal dynamics of Daphnia pulex in fish and fishless
ponds. MS Thesis. Michigan State University.

Gerrodette, T. 1987. A power analysis for detecting trends. Ecology
88(5): 1364-1372.

Gilbert, J. J. 1985. Competition between rotifers and Daphnia. Ecology
66(6): 1943-1950.

---. 1988a. Suppression of rotifer populations by Daphnia: a review of the
evidence, the mechanisms, and the effects on zooplankton community
structure. Limnology and Oceanography 33(6): 1286-1303.

---. 1988b. Susceptibilities of ten rotifer species to interference from
Daphnia pulex. Ecology 69(6): 1826-1838.

---. 1990. Differential effects of Anabaena aflinis on cladocerans and rotifers:
mechanisms and implications. Ecology 71(5): 1727-1740.

Gilbert, J. J ., and M. W. Durand. 1990. Effect of Anabaena flos-aquae on the
abilities of Daphnia and Keratella to feed and reproduce on unicellular algae.
Freshwater Biology 24:577-596.

Gilbert, J. J ., and K. L. Kirk. 1988. Escape responses of the rotifer Keratella:
Description, stimulation, fluid dynamics, and ecological significance.
Limnology and Oceanography 33(6): 1440-1450.

Gilbert, J. J., and R. S. Stemberger. 1985. Control of Keratella populations by
interference competition from Daphnia. Limnology and Oceanography
30(1):]80-188.

Gilbert, J. J ., and C. E. Williamson. 1978. Predator-prey behavior and its effect
on rotifer survival in associations of Mesocyclops edax, Asplanchna girodi,
Polyarthra vulgaris, and Keratella cochlean's. Oecologia 37: 13-22.

Gill, J. L. 1978. Design and analysis of experiments in the animal and medical
sciences, 1st Edition. Volume 3. The Iowa State University Press, Ames, IA.

Goulden, C.E., L. L. Henry and A. J. Tessier. 1982. Body size, energy

reserves and competitive ability in three species of Cladocera. Ecology
63: 1780-1789.

88

Hall, D. J., and T. J. Ehlinger. 1989. Perturbation, planktivory, and pelagic
community structure: the consequence of winterkill in a small lake. Canadian
Journal of Fisheries and Aquatic Sciences 46:2203-2209.

Hurlbert, S. H. 1984. Pseudoreplication and the design of ecological field
experiments. Ecological monographs 54:187-211.

Kerfoot, W. C., and W. R. DeMott. 1980. Foundations for evaluating community
interactions: the use of enclosures to investigate the coexistence of Daphnia
and Bosmina. Pages 725-741 in W. C. Kerfoot, editor. Ecology and Evolution
of Zooplankton communities. University Press of New England, Hanover,

N .H .

Kirk, K. L., and J. J. Gilbert. 1988. Escape behavior of Polyarthra in response to
artificial flow stimuli. Bulletin of Marine Sciences 43:551-560.

---. 1990. Suspended clay and the population dynamics of planktonic rotifers and
cladocerans. Ecology 71(5):1741-1755.

Koehl, M. A. R., and J. R. Strickler. 1981. Copepod feeding currents: food capture
at low Reynolds number. Limnology and Oceanography 26(6): 1062-1073.

Likens, G. E., and J. J. Gilbert. 1970. Notes on quantitative sampling of natural
populations of planktonic rotifers. Limnology and Oceanography 15:816-820.

Lind, O. T. 1979. Handbook of common methods in limnology, 2nd Edition.
Mosby, St. Louis.

Lynch, M. 1978. Complex interactions between natural coexploiters Daphnia and
Ceriodaphnia. Ecology 59(4):552-564.

---. 1979. Predation, competition, and zooplankton community structure: an
experimental study. Limnology and Oceanography 24(2):253-272.

MacIsaac, H. J., and J. J. Gilbert. 1989. Competition between rotifers and
cladocerans of different body sizes. Oecologia 81:295-301.

---. 1991a. Competition between Keratella cochlearis and Daphnia ambigua: effects
of temporal patterns of food availability. Freshwater Biology 25: 189-198.

---. 1991b. Discrimination between exploitative and interference competition
between Cladocera and Keratella cochlearis. Ecology 72:924-937.

89

Manny, B. A. 1971. Interactions of dissolved and particulate nitrogen in
lake metabolism. Thesis, Michigan State University, East Lansing, MI.

Manny, B. A., R. G. Wetzel, and R. E. Bailey. 1978. Paleolimnological
sedimentation of organic carbon, nitrogen, phosphorus, fossil pigments,
pollen, and diatoms in a hypereutrophic, hardwater lake: a case history
of eutrophication. Polskie Archiwum Hydrobiologii 25 :243-267.

McQueen, D. J ., J. R. Post, and E. L. Mills. 1986. Trophic relationships in
freshwater pelagic ecosystems. Canadian Journal of Fisheries and Aquatic
Sciences 43:1571-1581.

Neill, W. E. 1984. Regulation of rotifer densities by crustacean zooplankton in
an Oligotrophic montane lake in british columbia. Oecologia 61: 175-181.

Prepas, E. 1978. Sugar-frosted Daphnia: An improved fixation technique for
Cladocera. Limnology and Oceanography 23:557-559.

Scavia, D., G. A. Lang, and J. F. Kitchell. 1988. Dynamics of Lake Michigan
plankton: A model evaluation of nutrient loading, competition, and predation.
Canadian Journal of Fisheries and Aquatic Sciences 45: 165-177.

Schneider, D. W. 1990. Direct assessment of the independent effects of exploitative
and interference competition between Daphnia and rotifers. Limnology and
Oceanography 35(4):916-922.

Schoener, T. W. 1983. Field experiments on interspecific competition. The
American Naturalist 122:240-285.

Stemberger, R. S. 1979. A guide to rotifers of the laurentian great lakes.
U.S.E.P.A., Cincinnati.

Stemberger, R. S., and J. J. Gilbert. 1985. Body size, food concentration, and
population growth in planktonic rotifers. Ecology 66(4): 1 151-1 159.

Threlkeld, S. T. 1977. The midsummer dynamics of two daphnia species in
Wintergreen lake, michigan. PhD Thesis. Michigan State University.

Tilman, D. 1982. Resource competition and community structure. Princeton
University Press, Princeton, New Jersey.

Van Donk, E., R. D. Gulati, and M. P. Grimm. 1990. Restoration by
biomanipulation in a small hypertrophic lake: first-year results.
Hydrobiologia 191 :285-295.

90

Von Ende, C. N., and D. O. Dempsey. 1981. Apparent exclusion of the
cladoceran Bosmina longirostris by the invertebrate predator Chaoborus
americanus. The American Midland Naturalist 105:240-248.

Wetzel, R. G. 1983. Limnology. 2nd ed. Saunders College Publishing,
Philadelphia, PA

Wetzel, R. G., and G. E. Likens. 1991. Limnological Analyses, 2nd Edition.
Springer-Verlag, New York. )

Wickham, S. A., and J. J. Gilbert. 1991. Relative vulnerabilities of natural
rotifer and ciliate communities to cladocerans: laboratory and field
experiments. Freshwater Biology 26:77-86.

Wilkinson, L. 1990. Systat: the system for statistics. Systat, Inc., Evanston, IL.

Williamson, C. E. 1983. Invertebrate predation on planktonic rotifers.
Hydrobiologia 104:385-396.

Williamson, C. E., and N. M. Butler. 1986. Predation on rotifers by the

suspension-fading calanoid copepod Diaptomus pallidus. Limnology and
Oceanography 31(2):393-402.

91

"Illllllllllllllllllls