:1? .
3.25.5
u. .

I!
z 1| ‘
. L:
hm. .
‘whrbnrw.

Jr
.mmmfin
5:..- . .

mafia:

"
. . Hume.-

.: : t...
v

1.1.2:
.15—
...HW rm?!
, 5 3 ,
I r
z 15...: .
.. :: .. x

m3...

a4 w. .h"
En“- gfi
unti-

 

 

.

I R IV)“
14.7,... :
so.i‘i.1’iz... .
iflfihiifihnnua
4 23.12.?!»

.31.; :13.

TI

 

 

9
that ’1).
.43.. .1:

”are
hr.
can“

35%.

.p 3...:
.. 3.5.1.2.;
1:...

 

: v: f If?
.9 $12: A

.‘aL-t.
.5: a.
v4

 

3:93

 

THESIS
'7

5"

 

/7//!’*
,.LL:

kw

LIBRARY

Michigan State
University

 

 

 

This is to certify that the

thesis entitled
Epiphytic macroinvertebrates along a gradient of
Eurasian water milfoil (Myriophyllum spicatum L.):

The role of plant species and architecture

presented by

Kendra Spence Cheruvelil

has been accepted towards fulfillment
of the requirements for

M. S. degree inEisherieand Wildlife

Quagw

Major professor

Date 51/330! 00

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

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE 1 DATE DUE DATE DUE

 

NOV 1 37 Mn“ 4

 

Haiti‘s 82§5 l

 

was?“ 3’

 

 

W

:4. 23 u8

 

u

’0 MROZLQ 295

 

 

 

 

 

 

 

 

 

11m WM

EPIPHYTIC MACROINVERTEBRATES ALONG A GRADIENT OF EURASIAN
WATER MILFOIL (MYRIOPHYLLUM SPICA TUM L.): THE ROLE OF PLANT
SPECIES AND ARCHITECTURE
By

Kendra Spence Cheruvelil

A THESIS

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

MASTER OF SCIENCE

Department of Fisheries and Wildlife

2000

m:

 

ABSTRACT
EPIPHYTIC MACROINVERTEBRATES ALONG A GRADIENT OF EURASIAN
WATER MILFOIL (MYRIOPHYLLUM SPICATUM L.): THE ROLE OF PLANT
SPECIES AND ARCHITECTURE
BY

KENDRA SPENCE CHERUVELIL

An important feature of lake foodwebs are the interactions between submerged
macrophytes, macroinvertebrates, and the spread of the exotic macrophyte Eurasian water
milfoil (hereafter milfoil, Myriophyllum spicatum L.). To examine these interactions, I
had the following five objectives: 1) Design a sampler to sample macroinvertebrates
associated with submerged plants; 2) Assess the sample size and statistical power to
detect difl‘erences in macroinvertebrate density and biomass among plant species and
architecture types; 3) Examine macroinvertebrate density and biomass on different
species and architecture types of submerged plants; 4) Use meta-analysis to quantitatively
synthesize the published literature on the relationship between macroinvertebrates and
plant architecture; and 5) Examine patterns between macroinvertebrate density and
biomass and percent cover of milfoil. I designed and used a mesh bag sampler to sample
epiphytic macroinvertebrates fi'om one lake in 1998 and six lakes in 1999, all located in
southern MI. To plan my 1999 sampling, I used results from 1998 that indicated high
power to detect differences between two plant architecture types associated with a
moderate number of samples. Based on the 1999 field study and meta-analysis, I found
higher macroinvertebrate densities and biomass on dissected plants than undissected
plants. I also found that as percent milfoil increased, macroinvertebrate density and
biomass decreased along the six-lake gradient.

 

 

Rat

the

her

“P

CO]

Ell

[hi

ml

or

DEDICATION
There are many, many special people who have helped me make this degree possible.
Rather than name them all (and accidentally leave someone out!), I wish to dedicate this
thesis to two pivotal people. First, my mother. Noel Lynne Hogarth Spence sacrificed
her own Master’s degree to marry my father and bring me into the world, and later gave
up her demanding job to stay at home with me. I have learned so much from her, and
continue to do so. Her dream that I get my MS. has finally come true, and I owe her a
lot. The second person I dedicate this piece of work to is my husband, Jubin Joseph
Cheruvelil. He gave up everything he had and everyone he knew to marry me, and had
enough faith for both of us that we would be happy and successful in Michigan. Through
the last three years he has done much more than understand late nights and grumpy
moods - he has encouraged and supported me every step of the way. Mom, Jubin -— this

one’s for you.

ACKNOWLEDGEMENTS

This project was supported by funds from the US. E.P.A. S.T.A.R. Fellowship Program,
Michigan State University Agricultural Experimental Station, and Michigan State
University Extension. Thanks to John Madsen, Kurt Getsinger, Mike Stewart, and Chetta
Owens with the US. Army Engineer Research and Development Center for sharing their
macrophyte survey data and aiding in plant identification. Thanks to Helene Cyr and John
Downing for sharing their raw data for my meta-analysis. Special thanks to my advisor,
Patricia Soranno, and my committee members, Richard Merritt, Mary Bremigan, and
John Madsen for their helpful guidance and constructive consideration throughout the
past three years. Kristy Rogers, Steve Hanson, Tamara Brunger, Marla Sanbom, Sarah
Walsh, Rebekah Serbin, Abby Mahan, Vern Moore, Jason Stockwell, George Klemolin,
and Ray Valley provided invaluable sample collection and processing. I wish to thank
Marla Sanbom, in particular, for her enormous help processing macroinvertebrate
samples, conducting meta-analysis literature searches, and reading early thesis drafts.
Rich Merritt, Ace Samelle, Mary Bremigan, Sarah Walsh, Stacy Nelson, Ray Valley,
Nancy Nate, and Marla Sanbom provided me with many helpful reviews of earlier drafts.
Finally, I would like to recognize my partner in crime, Sarah Walsh, for getting me

through the hard days and helping to making the last three years enjoyable.

iv

PREFACE

Chapters 1 and 2 are written as manuscripts that either has been (Chapter 1) or will soon
be (Chapter 2) submitted. I would like to acknowledge my co-authors for their

contributions.

Chapter 1:
Cheruvelil, K.S., P.A. Soranno, and RD. Serbin. Macroinvertebrates associated with
submerged macrophytes: Sample size and power to detect effects. Submitted to

Hydrobiologia February 2000.

Chapter 2:

Cheruvelil, K.S., P.A. Soranno, J .D. Madsen, and MJ. Sanbom. The effects of Eurasian
water milfoil invasion on interactions between plant architecture and epiphytic
macroinvertebrates. To be submitted to Canadian Journal of Fisheries and

Aquatic Sciences August 2000.

 

 

Lis

Lis

TABLE OF CONTENTS

List of Tables ....................................................................................... viii
List of Figures ......................................................................................... x
Introduction ............................................................................................ 1
Interactions among macrophytes, macroinvertebrates, and fish ....................... 1
Macrophytes and epiphytic macroinvertebrates ......................................... 3
Eurasian water milfoil ....................................................................... 7
A whole-lake ecosystem experiment .................................................... 10

Chapter 1: Macroinvertebrates associated with submerged macrophytes: Sample size

and power to detect effects ........................................................................ 14
Introduction ................................................................................. 14
Materials and Methods ..................................................................... 16

Study Site ........................................................................... 16
Sampling ........................................................................... l7
Sampler description ...................................................... 17

Sample protocol .......................................................... 18

Sample processing ....................................................... 18

Data analysis ....................................................................... 19
Results ....................................................................................... 19
Discussion and Conclusion ............................................................... 20

Chapter 2: Interactions between plant architecture and epiphytic macroinvertebrates

and the effects of Eurasian water milfoil invasion ............................................... 23
Introduction ................................................................................. 23
Materials and Methods ..................................................................... 27

Study Area ......................................................................... 27
Sampling ........................................................................... 27
Macrophytes ............................................................... 27
Lake characteristics ...................................................... 27
Macroinvertebrates ...................................................... 28

Data analysis ....................................................................... 29
Macrophytes ............................................................... 29
Macroinvertebrates ...................................................... 3 1
Meta-Analysis ..................................................................... 32
Results ....................................................................................... 36
Lake characteristics ............................................................... 36
Macroinvertebrates, plant species, and plant architecture ................... 37
Pooled lakes ............................................................... 37

Within and across lakes ................................................. 38

 

 

Ap

Ap

AP:

Lite

Meta-analysis ............................................................. 38

Macroinvertebrates along a percent milfoil gradient ......................... 39
Discussion and Conclusion ............................................................... 4O
Macroinvertebrates, plant species, and plant architecture ................... 4O
Macroinvertebrates along a percent milfoil gradient ......................... 43
Conclusions/ Management Implications ......................................................... 49
Appendix 1: Chapter 1 Tables and Figures ...................................................... 56
Appendix 2: Chapter 2 Tables and Figures ...................................................... 64
Appendix 3: Taxa codes, taxa, and functional feeding groups ................................ 89
Appendix 4: Raw data: macroinvertebrate density .............................................. 91
Appendix 5: Raw data: macroinvertebrate biomass ........................................... 101
Appendix 6: Raw data: macroinvertebrate length ............................................. 111
Appendix 7: Raw data: macrophytes ............................................................ 121
Appendix 8: Aggregated raw data: macroinvertebrates per plant stem .................... 125
Literature Cited .................................................................................... 128

 

 

 

 

T2

T:

I:

T:

LIST OF TABLES

Chapter 1:

Table 1.

Examples of previous studies examining epiphytic macroinvertebrates in individual
lakes.

Chapter 2:

Table 1.

Lake macrophyte characteristics. The percent-dissected plants sampled
indicates the percent of plants sampled for epiphytic macroinvertebrates that
had dissected-leaves. Lakes are numbered to correspond with numbers on

Figure 1.

Table 2. Studies included in the meta-anlyses. The number of times sampled is the

Table 3.

number of times organisms were sampled within one season. The asterisk (*)
indicates that the three ponds included plant species of only one architecture
type, so we combined the three lakes for the meta-analyses.

Lake characteristics. TP stands for total phosphorous and TN stands for total
nitrogen. Lakes are numbered to correspond with numbers on Figure 1 and the
littoral zone is defined as the area from shore to the deepest point at which
plants consistently occur.

Table 4. The six most dominant macroinvertebrate taxa (total 2 70% of total

macroinvertebrate density and biomass) by macroinvertebrate density and
biomass and their associated functional feeding group across all six lakes.
Percentages refer to the percentage those six dominant taxa make up of the total
taxa.

Table 5. p-value, power, and number of plant species that needed to be sampled to detect

differences between dissected and undissected plants with an alpha of 0.05
within each lake. Numbers in parentheses indicate average number of plant
species within architecture groups actually sampled. Bold numbers indicate high
power (> 0.7) and asterisks indicate significant differences between dissected
and undissected plants at alpha 3 0.05. Lakes are numbered to correspond with
numbers on Figure 1.

Table 6. Adjusted Bonferonni p-values from an AN OVA test for differences between
mean macroinvertebrate densities on dissected versus undissected plants for all
six lakes in July and August. Bolded numbers represent values significant at
alpha < 0.05. Lakes are numbered to correspond with numbers on Figure 1.

Table 7. Adjusted Bonferonni p-values fiom an AN OVA test for differences between
mean macroinvertebrate biomass on dissected versus undissected plants for all 6
lakes in July and August. Bolded numbers represent values significant at alpha
< 0.05. Lakes are numbered to correspond with numbers on Figure 1.

LIST OF FIGURES

Chapter 1:

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Epiphytic macroinvertebrate mesh bag sampler that is a modification of the
folding quadrat sampler (Welch 1948). The sampler has the dimensions of 65
X 24 cm and is constructed from 200 um and 500 um mesh, 2 steel rings, and
canvas. It is closed at the bottom by a drawstring.

Five common macrophyte species of Heron Lake, MI, USA. Undissected: a)
P. illinoensis and b) P. richardsonii; Dissected: c) P. pectinatus, d) M.
spicatum, and e) Ranunculus Sp. Adapted from Fassett (1957).

The relationship between the number of samples and power at alpha = 0.05 and
effect size = 0.872. The number of samples necessary are indicated by circles
for plant architecture and triangles are for plant species.

The relationship between the number of samples and effect size for a) plant
species and b) plant architecture (alpha = 0.05). Power levels shown are 0.9
(circles), 0.8 (triangles), and 0.5 (squares). For comparison, the lower two
graphs show the enlarged region of 0-50 samples.

Macroinvertebrate a) density and b) biomass by plant species and architecture.
Plant species are abbreviated as: 111 (P. illinoensis), Ric (P. richardsonii), Pec
(P. pectinatus), Spi (M. spicatum), and Ran (Ranunculus sp.). Bars represent
the standard error for each plant species.

Chapter 2:

Figure 1. Map of Michigan counties and study lakes. 1 = Camp Lake, 2 = Big Crooked

Lake, 3 = Lobdell Lake, 4 = Heron Lake, 5 = Clear Lake, 6 = Big Seven Lake.

Figure 2. Weighted frequency of plant species in each lake in the vegetated littoral zone

August 1999 (J .D. Madsen unpublished data). Lakes are in order fiom low
percent milfoil cover (1) to high percent milfoil cover (6). Full scientific
names are: Cabomba caroliana Gray, Ceratophyllum demersum L., Elodea
canadensis Michx., Heteranthera dubia Jacq., Myriophyllum spicatum L.,
Najas spp., Potamogeton amplifolius Tuckerm., Potamogeton crispus L.,
Potamogeton foliosus G., Potamogeton gramineus L., Potamogeton illinoensis
Morong., Potamogeton natans L., Potamogeton nodosus Poir., Potamogeton

Figure 3.

Figure 4.

Figure 5.

Figure 6.

pectinatus L., Potamogeton praelongus Wulf., Potamogeton pusillus L.,
Potamogeton richardsonii Benn., Potamogeton robbinsii Oakes., Potamogeton
strictifolius Benn., Potamogeton sp., Potamogeton zosteriformis Fernald.,
Ranunculus sp., Utricularia spp., Valisneria americana Michx, Zannichellia
sp. An asterisk (*) indicates plant species that were sampled for epiphytic
macroinvertebrates.

The functional feeding groups by macroinvertebrate density (#) and biomass
(mg) with the six lakes pooled. Percentages refer to the percentage each
functional feeding group makes up of the total macroinvertebrate density or
biomass in July and August, respectively.

Average macroinvertebrate density (#) and biomass (mg) per g plant biomass
within functional feeding groups in July (A and C) and August (B and D).

Data were analyzed by pooling lakes. Bars represent the standard error for each
functional feeding group. F. and G. collector stands for filtering and gathering
collectors, respectively.

Average macroinvertebrate density (#) (A and B) and biomass (mg) (C and D)
per g plant biomass for the three (July) and four (August) plant species that
were sampled in 3 - 6 of the lakes. Data were analyzed by pooling lakes. Large
circles represent plant species not sampled. Plant species that are italicized are
dissected and plant species that are not italicized are undissected. Bars
represent the standard error for each plant species.

Average macroinvertebrate density (#) (A and B) and biomass (mg) (C and D)
per g plant biomass for the two plant architecture groups. Data were analyzed
by pooling lakes. Bars represent the standard error for each architecture type.

Figure 7. Effect size (natural log response ratio) for each study included in the

unweighted meta-analysis (A) and the weighted meta-analysis (B). An effect
size greater than zero means that dissected plants exhibit higher
macroinvertebrate densities than undissected plants. Filled diamonds represent
the 18 studies included in the unweighted meta-analysis, empty circles
represent the six individual lakes in this study, the filled cross represents the
pooled six lakes fiom this study (not included in the weighted meta-analysis)
and the filled square is the mean effect size (calculated as the weighted and
unweighted natural log response ratio of the mean macroinvertebrate density
on dissected plants/ mean macroinvertebrate density on undissected plants).
An asterisk (*) indicates those studies that averaged macroinvertebrate density
across multiple lakes. 1 refers to the pooled six lakes in this study, which is

 

 

Figure

Figure

Figur

Figut

Figu

Figt

Cor

Fig:

shown for comparison only and was not included in the weighted meta-
analysis. Memph. is short for Memphremagog.

Figure 8. The proportion of dissected plants present across lakes along the weighted
percent milfoil gradient in August (J.D. Madsen unpublished data).

Figure 9. Average macroinvertebrate density (#) and biomass (mg) per g plant biomass
along the weighted percent milfoil gradient in July (A and C) and August (B
and D). Data were analyzed across lakes and each data points represent the
average density or biomass for each plant species within each lake.

Figure 10. Average density (#) and biomass (mg) of odonates and chironomids per g
plant biomass in Camp Lake in July and August.

Figure 11. Total macroinvertebrate density (#) and biomass (mg) per g plant biomass (A
and B) and total macroinvertebrate density (#) and biomass (mg) per m 2
vegetated littoral zone (C and D) across lakes along the weighted percent
milfoil gradient in July and August.

Figure 12. Total macroinvertebrate density (#) and biomass associated with milfoil per g
plant biomass across lakes along the weighted percent milfoil gradient in July
(A and C) and August (B and D).

Figure 13. Cumulative summer (July and August) bluegill densities (#) per m2 across
lakes along the weighted percent milfoil gradient (R.D. Valley unpublished
data). No data was collected from Camp Lake.

Conclusions/Management Recommendations:

Figure 1. Average macroinvertebrate density (#) and biomass (mg) per g plant biomass
in reference and treatment lakes during July (A and D), August (B and E), and
August excluding Camp Lake (C and F). Bars represent the standard error for
each architecture type.

indit
bent
orga
lags
mac
imp.
con
00‘
the ‘
whi

Inle

"if“

INTRODUCTION

To fully understand lake ecosystems, ecologists must concentrate not on
individual ecosystem components, but examine the many complex inter-relationships
between components. Complex interactions result from multiple pathways linking
organisms and abiotic resources, and involve both direct and indirect effects and time
lags (Carpenter 1988). An example of a complex lake interaction occurs among
macrophytes, macroinvertebrates, fish, and humans. Epiphytic macroinvertebrates are an
important forage base for many species of juvenile fish that use macrophyte beds for
cover and as a source for food (Keast 1984; Diehl and Kornijow1998; Persson and
Crowder 1998). However, macrophytes are diverse in shape and form, and their role in
the food web is dependent on their abundance and community composition, both of
which are affected by human management practices (Olson et a1. 1998).

Interactions among macrophytes, macroin vertebrates, and fish

Specifically, the interaction between juvenile bluegill sunfish (Lepomis
macrochirus) and its major predator, the largemouth bass (Micropterus salmoides)
depends on macrophyte populations. Adult bluegills make use of macrophytes for cover
and make daily migrations from the pelagic zone where they feed on zooplankton, to the
edges of the littoral zone where they are more protected from predators (Werner and Hall
1988). Juvenile bluegills also use macrophytes for cover and food. They feed on
epiphytic macroinvertebrates within the vegetated littoral zone where they compete with
young-of-year largemouth bass for epiphytic macroinvertebrates until bass reach a
sufficiently large size to make the switch to the more energetically profitable fish diet

(Olson et a1. 1995).

Questions still surround this macrophyte-mediated interaction between
macroinvertebrates and fish. Mittelbach (1988) experimentally studied the effects of fish
on macroinvertebrates in a natural lake, where predator and prey have co-occurred for
many generations. He found that the fish in this lake had strong effects on
macroinvertebrate size structure, little or no effect on species richness, and variable
effects on total macroinvertebrate densities (Mittelbach 1988). Macroinvertebrates, in
turn, can affect fish abundance and growth. Crowder and Cooper (1982) found that
bluegill in experimental ponds with intermediate macrophyte density exhibited higher
growth than bluegill in low or high plant density ponds because at low macrophyte
densities there was insufficient food and at high densities fish search and capture times
were long. However, laboratory and mesocosm studies show that plant density does not
influence bluegill growth beyond relatively low densities owing to the interaction
between capture probabilities and macroinvertebrate densities (Savino et a1. 1992).
Therefore, high plant densities may have either no effect or a negative effect on bluegill
growth, which is driven by both macrophyte and macroinvertebrate densities.

Most studies examining interactions among fish and macroinvertebrate prey have
either considered the effects of introduced fish in a previously fishless system (Crowder
and Cooper 1982), or were conducted in artificial ponds (Savino et a1. 1992). In addition,
most studies concentrate on the effects of macrophyte density or biomass only, even
though it has been suggested that macrophyte species composition, architecture, and
growth form all have effects on fish foraging (Dionne and Folt 1991; Dibble and Harrel

1997). We know surprisingly little about the potential effects of submerged macrophyte

 

 

 

 

speci
andl

llaa

usui
Vruhi
flout
enha
Lodg
grou
Spec
such
and
habi
Pho:

Lod

Phy

[eve

species composition, architecture, and growth form on macroinvertebrate colonization,
and how these effects are translated up the food chain to fish.
Macrophytes and epiphytic macroinvertebrates

Macrophytes, the largest sessile organism in fresh water ecosystems, have large
effects on physical, biogeochemical, and biotic habitats (Carpenter and Lodge 1986;
Engel 1990; Barko and James 1998). The most important physical effects of macrophyte
structure are those on 1) light: extinction coefficients vary among plant species and
significantly alter the depth profile of photosynthesis, 2) temperature: vertical gradients
within macrophyte stands are much steeper than neighboring unvegetated areas, 3) water
flow: flow is reduced among macrophyte beds, and 4) substrate: macrophyte beds
enhance deposition of fine sediments that would otherwise be eroded (Carpenter and
Lodge 1986; Engel 1990; Barko and James 1998). Biogeocherrrically, macrophyte
growth alters 1) diel and annual oxygen dynamics, 2) dissolved inorganic carbon
speciation and pH, 3) dissolved organic carbon levels, and 4) dissolved nutrient levels
such as the limiting nutrient, phosphorus (Carpenter and Lodge 1986; Engel 1990; Barko
A and James 1998). In addition, macrophyte decomposition changes the biogeochemical
habitat through 1) the release of dissolved substances such as organic carbon,
phosphorus, and nitrogen, 2) deoxygenation, and 3) sediment accretion (Carpenter and
Lodge 1986; Engel 1990; Barko and James 1998).

The effects of macrophytes on biotic interactions are dependent upon macrophyte
physical structure, biomass and productivity, which all vary substantially within and
among lakes (Crowder and Cooper 1982; Carpenter and Lodge 1986; Engel 1990). The

level of productivity drives how much organic carbon macrophytes release, which in turn

 

prov
diss:
mac
thes
our

199l

pan
Sub
mor

Lill;

mac
bee:
km
and
occ
ther
1112K
bett
198
hab

dele

provides a substrate for periphyton and macroinvertebrates. In turn, periphyton exchange
dissolved nutrients with the water, assimilate phosphorus released from decomposing
macrophytes, and provide food for many macrophyte-associated grazers. Commonly
these macroinvertebrate grazers use macrophytes for food, refuge from predation,
oviposition sites, and access to the air-water interface (Carpenter and Lodge 1986; Engel
1990; Newman 1991; Merritt and Cumrnins 1996).

The relationship between macroinvertebrates and submerged macrophytes can be
partly explained by macrophyte physical structure, also known as architecture.
Submerged macrophytes can be grouped according to architecture based on plant
morphology (the number, morphometry, and arrangement of stems, branches, and leaves;
Lillie and Budd 1992). Macrophytes are diverse in shape and form, and the architecture
of plants has been suggested to influence the importance and colonization of epiphytic
macroinvertebrates (Jackson 1997). Specifically, plants with finely dissected leaves have
been found to support more macroinvertebrates than plants with broader, undissected
leaves (Krecker 1939; Andrews and Hasler 1943; Gerking 1957; Mrachek 1966; Gerrish
and Bristow 1979; Cattaneo and Kalff 1980; Dvorak and Best 1982). This pattern may
occur because dissected-leaf plants have a higher surface area to volume ratio and
therefore provide more habitat for macroinvertebrate colonization, more food for grazing
macroinvertebrates in the form of periphyton, or additional complexity which offers
better refirge from predators (Krull 1970; Pardue 1973; Dvorak and Best 1982; Gilinsky
1984; Jackson 1997). In fact, with surface area held constant, J effries (1993) found that
habitat complexity (measured by fractal dimensions) is an important factor in

determining invertebrate densities on macrophytes of differing morphologies.

 

 

19

f0‘

1e:

Sll

Whereas many studies have found that dissected-leaf plants support more
macroinvertebrates than undissected-leaf plants (Krecker 1939; Andrews and Hasler
1943; Gerking 1957; Mrachek 1966; Gerrish and Bristow 1979), a more recent study
found that macroinvertebrate density did not vary predictably with leaf dissection across
multiple lakes (Cyr and Downing 1988a and b). Instead, macroinvertebrate density was
plant species-specific and related to plant biomass rather than dissection (Cyr and
Downing 1988a and b). In their study of high surface area, finely divided, and thinly
leafed plants, Parsons and Matthews (1995) suggest that the nature of the colonizable
surface (soft or hard stems, brittle or pliable leaves, whorled or non-whorled leaves) is at
least as important as the surface area. In addition, a study relating biomass and surface
area of six submerged plants did not find that all dissected-leaf plants had higher surface
areas than plants with undissected leaves (Sher-Kan] et al. 1995). This lack of consensus
has contributed to the many questions that remain regarding the macroinvertebrate-
macrophyte relationship.

Because macroinvertebrate densities are much higher in vegetated areas than
open—water areas, these organisms are an extremely important component of lake food
webs (Gerking 1957; McLachlan 1969; Krull 1970; Biggs and Malthus 1982; Watkins et
a1. 1983; Pardue and Webb 1985; Engel 1988; Jackson 1997). Although we recognize
the importance of epiphytic macroinvertebrates and the macrophytes they colonize, the
relationship between macrophytes and epiphytic macroinvertebrates, and the roles these
organisms play in lake food webs has been difficult to quantify (Downing and Cyr 1985),
leading to uncertainty and misinterpretation of the importance of epiphytic

macroinvertebrates in lake ecosystems. This uncertainty is partly because sampling

macroinvertebrates on submerged plants is difficult. Many sampling methods are
expensive, cumbersome, time consuming, and often require trained SCUBA personnel,
which can limit the number of replicates taken (Mittelbach 1981a; Downing 1986; Creed
and Sheldon 1992; Galanti 1995). Other methods are semi-quantitative because the
sampling method disturbs the plants causing a loss of organisms that does not occur
consistently and thus cannot be quantified (e.g. Krecker 1939; Rosine 1955; Schramm
and Jirka 1989; Beckett et a1. 1992). Also, many samplers combine benthic and epiphytic
macroinvertebrates, not allowing for the separation of organisms between these two very
different habitats (e.g. Hanson 1990; Hargeby 1990; James et al. 1998). In addition to
these sampling issues, most studies have sampled only one lake, a small number of plant
species, or have not used comparable methods to sample, process, analyze, and report
data (e. g. macroinvertebrates expressed as density per plant, density per plant biomass,
density per plant surface area, density per unit bottom surface, or density per unit water
volume; Downing 1984; Kornikova 1971; Downing and Cyr 1985; Jackson 1997).
Therefore, it is difficult to determine what factor(s) contribute to the different study
conclusions.

To further complicate the study of epiphytic macroinvertebrates, these organisms
exhibit large plant-to-plant variability. This variability has been attributed to predation,
periodic macroinvertebrate emergence, irregular plant density, irregular plant species
distribution, seasonal plant succession, fluctuations in macroinvertebrate food supply,
appearance of new macroinvertebrate broods, natural mortality, and occurrence of
macroinvertebrates of the same species but of different size (Gaufin et a1. 1956;

Mracheck 1966; Soszka 1975). This variability makes replication important, especially

 

re}

Jo?

ca'

m2

sul

 

int

sta

0r

ca]

8))

pr:

19

He

(11:

Fr

when population densities are low or a small sampler is being used (Resh 1979; Downing
and Cyr 1985). Although estimates of statistical power in ecological studies have been
reported in some recent studies (e. g. Chick and McIvor 1994; Carpenter et al. 19953;
Johnson 1998), these important and biologically relevant statistics are still too seldom
calculated and, in particular, have not been examined for studies of epiphytic
macroinvertebrates.
Eurasian water milfoil

Eurasian water milfoil (hereafter rrrilfoil, Myriophyllum spicatum L.) is an exotic
submerged macrophyte found in much of temperate North America. Milfoil was
introduced to North America prior to 1950 from Europe and by 1985 was reported in 33
states, the District of Columbia, and the Canadian Provinces of British Columbia,
Ontario, and Quebec (Couch and Nelson 1985). Milfoil forms dense surface mats, or
canopies, that suppress native plant growth, and lead to homogeneous macrophyte beds
(Aiken et al. 197 9; Madsen et al. 1988), in addition to interfering with recreational
swimming and boating (Newroth 1985). Because milfoil has three mechanisms of
propagation, (seed production, stolon formation, and fiagmentation; Madsen and Smith
1997) and can grow in water from 1 - 10 m deep, it has spread rapidly throughout North
America (Aiken et al. 1979). Seeds serve as long-term mechanisms of reproduction,
enabling the species to survive protracted periods of dormancy (Madsen 1991). Stolons
(stems that form adventitious roots) extend outward from the parent plant and produce
new plants in the immediate area, thus allowing populations to disperse locally over
distances of a few meters or less (Madsen et al. 1988; Madsen and Smith 1997).

Fragmentation, another type of vegetative clonal propagation, is the predominant means

 

 

of d
that
gen
the

han

pror
198
mac
194;
Dvo
area
pect
Hou
inve.
EVer
bioir
areas
he be
to Sn}
mart

mOVe

of dispersal over longer distances (Madsen et al. 1988). The two types of fragmentation
that milfoil exhibit are 1) autofragmentation - self-induced abscission of shoot apices,
generally after attaining peak biomass, and 2) allofragmentation - mechanical breakage of
the plant stem by disturbances in the water, such as those generated by mechanical
harvesters, boats, swimmers, animals, and wave action (Madsen and Smith 1997). These
three mechanisms of propagation allow milfoil to spread quickly.

Because macrophyte structural complexity is species-specific, certain species
provide more substrate for macroinvertebrates and cover for fish (Cyr and Downing
1988a). Milfoil is a dissected-leaf plant, and dissected-leaf plants usually have higher
macroinvertebrate densities associated with them (Krecker 1939; Andrews and Hasler
1943; Gerking 1957; Mrachek 1966; Gerrish and Bristow 1979; Cattaneo and Kalff 1980;
Dvorak and Best 1982). For the same unit of biomass, milfoil also has a higher surface
area than four other native plant species (N. obtuse, P. lucens, P. perfoliatus, P.
pectinatus; Sher-Kan] et al. 1995), and a low frequency of interstices (Dibble et al. 1996).
However, contrary to these findings, rrrilfoil has been found to support fewer
invertebrates than native plant species (Soszka 1975; Keast 1984; Cattaneo et al.1998).
Even within a milfoil bed, macroinvertebrate density differs. Macroinvertebrate density,
biomass, and taxa richness is higher in the upper and edge areas than lower and center
areas (Sloey et al. 1997). Low macroinvertebrate densities associated with milfoil may
be because milfoil forms dense, homogeneous vegetation beds, which have been shown
to support lower densities of epiphytic macroinvertebrates (Brown et al. 1988). Dense,
mat-forming, homogeneous milfoil beds may also reduce diurnal and seasonal fish

movement by acting as a barrier (Crowder and Cooper 1982; Keast 1984; Trebitz et al.

1996). In fact, Lyons (1989) has attributed a decline in small littoral zone fish species
diversity in Lake Mendota in part due to the invasion of milfoil. Thus, excessive milfoil
growth affects both the fish forage base and habitat, which can result in large disruptions
in littoral zone food webs.

There have been some reports of natural milfoil declines (Lake Wingra and
southern Ontario lakes; Carpenter 1980; Painter and McCabe 1988; Trebitz et al. 1993),
but the continued spread of this exotic species, and the possible recreational and
ecological ramifications of its spread, have prompted much research into milfoil ecology,
biology, and management (e.g., Keast 1984; Madsen et al. 1988; Smith and Barko 1990;
Chilton 1990; Trebitz et al. 1993; Madsen and Smith 1997). For example, there have
been many studies on the relative importance of milfoil’s various methods of
reproduction to its regional expansion and the implications of those reproductive
strategies for milfoil management (Madsen et al. 1988; Madsen and Smith 1997). Due to
its multiple propagation mechanisms, traditional management tools such as harvesting
without plant removal, derooting, dredging, and drawdown can actually promote
expansion of milfoil (Cooke et al. 1990; Smith and Barko 1990). To date, studies have
not provided managers with a clear framework for managing milfoil for the combined
purposes of fisheries, recreation, and water quality, all of which are negatively impacted
by milfoil invasions (Keast 1984; Newroth 1985; Carpenter and Lodge 1986).

A potential management option for controlling milfoil is the use of an aquatic
herbicide that is selective for milfoil. Sonar® (active ingredient fluridone, SePRO
Corporation, Indianapolis, IN) is a candidate for such an approach. Relative to most

native aquatic plant species, milfoil is highly susceptible to low concentrations of

 

 

 

flur
199
littl
trea
con
efft
mac

but

indi
Sor
the
pri\
stuc
trea
dire
Com

Aw

belt!
Ofp:
(Nit:

Scale

fluridone, increasing the potential for selective plant control (N etherland and Getsinger
1993). Sonar® is relatively non-toxic but must be applied to the whole lake. However,
little is known about the direct and indirect effects of this type of whole lake herbicide
treatment on the native macrophyte communities, the associated macroinvertebrate
communities, and the subsequent effects on fish populations. Past studies of the indirect
effects of Sonar® in two Minnesota lakes reported negative impacts on water quality,
macroinvertebrates (Delong and Mundahl 1996), and small littoral zone fish diversity,
but positive effects on larger fish growth (Pothoven 1996; Pothoven 1999).

As a result of remaining questions and insufficient data regarding the direct and
indirect effects of Sonar® treatments, many states, including Michigan, have restricted
Sonar® use. In May 1997, the Michigan Department of Environmental Quality (DEQ),
the US. Army Corps of Engineers (U SAE), Michigan State University (MSU), two
private environmental consulting firms, and a lake management interest group began a
study to determine the direct and indirect effects of whole-lake, low-dose Sonar®
treatments (5 ppb) on plant, fish, and invertebrate communities. The USAE examined the
direct effects of Sonar® treatments on macrophytes, while the MSU research group
continues to examine the indirect effects of Sonar® on invertebrate and fish populations.
A whole-lake ecosystem experiment

An ecosystem approach is most relevant to study the inherently complex linkages
between macrophytes, invertebrates and fish. In the past, there have been many studies
of particular macrophyte-mediated processes conducted at scales less than whole-lake
(Nichols and Keeney 1973; Dale and Gillespie 1977; Carpenter and Adams 1979), small-

scale studies of the consequences of macrophyte management, especially by herbicides

10

(Shel
ecosy
Nibb
studi
and 1
ma
and l
oppu
lakes

tropl

of th
1995
not 1
ecos
and ‘
eXpe
l’CSp.
ftp“
subj

Vari.

Sens

(Sheldon 1986; Netherland and Getsinger 1993), and modeling exercises examining
ecosystem responses to macrophyte manipulations (Mittelbach 1981b; Trebitz and
Nibbelink 1996). Although many important hypotheses have emerged from these
studies, the interactions among macrophytes and their management, macroinvertebrates,
and fish cannot be firlly understood without whole-lake experiments in which
macrophytes are manipulated and the response of the ecosystem is measured (Carpenter
and Lodge 1986, Oslon et al. 1998). The Michigan Sonar® project presented a unique
opportunity to participate in such an experiment, using multiple treatment and reference
lakes to examine the direct and indirect effects of an herbicide treatment on multiple lake
trophic levels.

Whereas ecosystem studies provide the opportunity to study real systems with all
of their intrinsic complexity under realistic spatial and temporal scales (Carpenter et al.
1995b; Carpenter 1996), that same complexity means all of the important processes may
not be measured or even detected. This complication can make the interpretation of
ecosystem study results difficult, unless all competing hypotheses have been identified
and tested separately. In many cases, conclusions are based on inference and the
experiment may not conclusively show that a particular process or mechanism is
responsible. Another drawback of ecosystem experiments is that it is often too costly to
replicate these studies and rigorous controls may not exist. The experiments are also
subject to variability beyond the control of the experimenter, such as the effects of
variable weather, which can complicate the interpretation of results (Carpenter 1989).

Because of these problems, ecosystem experiments do not have the same level of

sensitivity and precision of lab experiments, and often only a large response to a

11

treatment can be detected over natural variability (Carpenter 1989). Despite these
inherent drawbacks, freshwater ecosystem experiments have successfully been used to
address the responses of both communities and biogeochemical processes to a variety of
stresses (Carpenter et al. 1995b). Most recently, Olson et al. (1998) used a multi-lake
experiment to study the value of managing macrophytes to improve fish growth.

I designed my study on epiphytic macroinvertebrates keeping these potential
drawbacks of ecosystem experiments in mind. Using data from one lake sampled in
summer 1998, I estimated the statistical power and sample size necessary to detect
differences in macroinvertebrate density and biomass among plant species and
architectures. During summer 1999, I sampled six lakes, took approximately 75 replicate
samples within lakes, and pooled macroinvertebrate samples within plant species in
hopes of decreasing plant-to-plant variability so that I might better detect lake-to-lake
variability. I also recognized the potential confounding factors inherent in my study. For
example, treatment lakes in my study were subjected to additional plant management
strategies by riparian lake owners, such as other herbicide applications and mechanical
harvesting. Therefore, I cannot be certain that differences between reference and
treatment lakes are due to the Sonar® treatments alone. Thus, I study the relationships
and patterns within and among trophic levels along a gradient of percent milfoil cover,
with lakes low on the gradient as a result of Sonar® treatments. By recognizing and
compensating for some of the potential drawbacks of ecosystem experiments, I hope to
further explain the relationship between macroinvertebrates and macrophytes.

Many questions have yet to be answered regarding the interactions between

macrophytes and macroinvertebrates. We also know little about how the spread of the

12

exotic nrilfoil and our subsequent management actions affect those interactions. To

address some of these questions, I had the following five objectives:

1.

Design a sampler to sample macroinvertebrates associated with submerged
plants (Chapter 1).

Assess the sample size and statistical power to detect differences in
macroinvertebrate density and biomass among species of plants from
undissected and dissected plant architecture types using samples taken
from a single lake in August 1998 (Chapter 1).

Examine patterns between macroinvertebrate density and biomass and
plant species and architecture using samples taken from six lakes during
the summer of 1999 (Chapter 2).

Use meta-analysis to quantitatively synthesize the published literature on
the relationship between macroinvertebrates and plant architecture
(Chapter 2).

Examine patterns between macroinvertebrate density and biomass and the
percent cover of milfoil using samples taken from six lakes during the

summer of 1999 (Chapter 2).

13

Chapter 1: Macroinvertebrates associated with submerged macrophytes:

Sample size and power to detect effects

Introduction

When planning and conducting ecological experiments, it is important to consider
how many samples are necessary to detect differences among treatments with acceptably
high statistical power. Clearly, the goal is to maximize power (1 — Beta, the probability
of correctly rejecting the null hypothesis) by rrrinimizing beta (the probability of making
a type H error or failing to reject a false null hypothesis) (Peterrnan 1990a). Thus, for an
experiment with low power, little confidence can be placed in a conclusion based on the
failure to reject a null hypothesis. Power can be calculated for different assumed effect
sizes (the magnitude of the change in the parameter of interest that can be detected by an
experiment calculated as the arithmetic difference between the expected value and the
observed value for the parameter of interest) (Cohen 1988). The experiment should not be
performed if the detectable effect size is larger than the effect size that is biologically or
economically important (Rotenberry and Wiens 1985). Through these calculations of
power, a researcher can determine the feasibility of a study and anticipate how many
samples are necessary to detect differences among treatments with various levels of
power, thus facilitating better experimental design.

Although estimates of statistical power in ecological studies have been reported in
some recent studies (e.g. Carpenter et al. 1995a; Johnson 1998), these important and
biologically relevant statistics are still too seldom calculated and, in particular, have not

been examined for studies of epiphytic macroinvertebrates. An analysis of statistical

14

power is especially important when studying epiphytic macroinvertebrates because these
organisms exhibit large plant-to-plant variability due to predation, periodic
macroinvertebrate emergence, fluctuations in macroinvertebrate food supply, appearance
of new macroinvertebrate broods, natural mortality, and the occurrence of
macroinvertebrates of the same species but of different size (Gaufin et al. 1956; Mrachek
1966; Soszka 1975).

Epiphytic macroinvertebrates and the macrophytes they colonize are ecologically
important components of lake ecosystems. In particular, epiphytic macroinvertebrates
are an important forage base for many species of juvenile fish that use macrophyte beds
for cover and as a source for food (Diehl and Kornijow 1998). However, macrophytes
are diverse in shape and form, and the morphology of the plants themselves has been
suggested to influence the importance and colonization of epiphytic macroinvertebrates
(Jackson 1997). Submerged macrophytes can be grouped according to architecture based
on plant morphology (the number, morphometry, and arrangement of stems, branches and
leaves) (Lillie and Budd 1992). Macrophyte architecture type has been found to explain
some of the variation in the density of macroinvertebrates, with plants having finely
dissected leaves supporting more macroinvertebrates than plants with broader,
undissected leaves (Krecker 1939; Andrews and Hasler 1943; Gerking 1957; Mrachek
1966; Gerrish and Bristow 1979; Kershner and Lodge 1990; J effries 1993). It has been
suggested that this pattern occurs because most dissected-leaf plants provide more habitat
for colonization, more epiphyton for grazing macroinvertebrates, or additional

complexity that offers better refuge from predators. However, a more recent study found

15

that macroinvertebrate density did not vary predictably with leaf dissection across
multiple lakes (Cyr and Downing 1988a).

The patterns of epiphytic macroinvertebrate communities and their role in lentic
food webs have been difficult to quantify, partly because sampling macroinvertebrates on
submerged plants is difficult and past studies have not used comparable methods to
sample, process, analyze, and report data (Downing and Cyr 1985; Jackson 1997). In
addition, power analyses have not been conducted in any study examining the patterns of
epiphytic macroinvertebrates. Thus, questions remain about the relationship between
epiphytic macroinvertebrates and macrophytes, and whether these organisms are too
variable to discern patterns of density and biomass.

To address these questions, I designed a mesh bag sampler that is a modification
of the folding quadrat sampler (Welch 1948) to sample macroinvertebrates associated
with submerged plants. I assessed the sample size and statistical power to detect
differences in macroinvertebrate density and biomass among species of plants fi'om broad
and dissected plant architecture types. I also examined patterns between
macroinvertebrate density and biomass and plant species and architecture types. I
hypothesized that broad-leaf plants would harbor fewer macroinvertebrates than

dissected-leaf plants.

Materials and Methods
Study Site
I sampled epiphytic macroinvertebrates on August 4 and 5, 1998 in Heron Lake,

located in Seven Lakes State Park in SE. Michigan, U.S.A (42.81N, 83.52W). The lake

16

has an extensive forested riparian zone and undergoes very little plant management,
except for occasional mechanical harvesting in localized areas surrounding the public
boat launch and beach. The surface area of Heron Lake is 53 ha and the mean depth is
3.5 m. Nearly 65% of the lake is littoral (littoral zone defined as average depth beyond
which no plant growth is observed; ~4.6 m). Nineteen plant species were recorded
during macrophyte surveys performed in August of 1998 (J .D. Madsen unpublished
data).
Sampling

Sampler Description: I sampled individual plant stems with a mesh bag sampler
that is a modification of the folding quadrat sampler (Welch 1948) (Figure 1). It is
constructed of 200 and 500 um mesh, thus the sampler collects organisms > 500 pm.
The sides are constructed of 200 um mesh for flexibility, ease of construction, and
sampler deployment. Two brass rings provide structure to the mesh bag (the top ring is
smaller than the bottom ring for easy inversion of the sampler). All seams are on the
outside of the sampler, allowing for a smooth inner surface. The sampler is 65 cm long
and 24 cm in diameter. It has a drawstring at the bottom to close the sampler and trap the
sampled macrophyte and its associated macroinvertebrates. A crew of three people
performs the sampling: one snorkeller collects samples and two people process the
samples in a boat. The snorkeller positions the sampler above a randomly chosen plant
and slowly (to limit disruption and subsequent loss of organisms) lowers it down until the
desired plant length is inside the sampler. Then the drawstring is pulled taut, the plant
stem is broken off at its base, and the sampler is brought to the surface. The processors in

the boat cutoff any additional plant material extending beyond the sampler. The sampler

l7

is then inverted and rinsed, and the contents (macrophyte, macroinvertebrates, and water)
are stored in a sealed plastic bag. Samples are kept cool and dark for further processing.

Sample Protocol: I sampled five common plant species that fit into the two plant
architecture groups. Two species were classified as undissected, or broad-leafed plants:
Potamogeton richardsonii Benn. (clasping-leaf pondweed), and Potamogeton illinoensis
Morong. (Illinois pondweed); and three as dissected-leafed: Ranunculus sp. (water
crowfoot), Potamogeton pectinatus L. (sago pondweed), and Myriophyllum spicatum L.
(Eurasian water nrilfoil) (Figure 2). I sampled epiphytic macroinvertebrates at three sites
separated by greater than 100 m. Each site was approximately 2 m deep (average depth of
littoral zone) and contained each of the five plant species. I randomly sampled five
macrophytes of each species from approximately a 10 m radius around an anchored boat
at each site, resulting in 15 individuals of each plant species totaling 75 samples. I chose
these numbers of plant species and replicates based on comparisons with previous studies
(Table 1).

Sample Processing: In the lab, I rinsed all individual macrophyte samples with
water to detach insects, then dried the plants at 105 °C for 48 hours and weighed them to
estimate plant biomass. Macroinvertebrates were preserved in 95% ethanol, counted, and
identified to the lowest taxonomic level possible (usually genus). Length-weight
equations from the literature were used to estimate macroinvertebrate biomass from body
lengths measured using an ocular micrometer (Rogers et al. 1977; Smock 1980; Meyer

1989; Burgherr and Meyer 1997; G.G. Mittelbach unpublished data).

18

Data Analysis

For all analyses, I standardized macroinvertebrate density and biomass (expressed
as numbers and mg of animals) by plant dry weight, which allows for the comparison of
macroinvertebrate density and biomass among different plant species and architecture
types. I conducted sample size and power analyses using PASS 6.0 software (NCSS
Statistical Software 1998). I calculated the power to detect differences in
macroinvertebrate density and biomass among the five plant species and two architecture
types given the number of samples taken. Using one-way ANOVAs and setting alpha =
0.05, I estimated the number of samples necessary to detect differences in
macroinvertebrate density and biomass between the five plant species and two
architecture types at different levels of power and a fixed effect size. I also calculated the
number of samples necessary to detect differences in macroinvertebrate density and
biomass among the five plant species and two architecture types with a fixed power level
and various effect sizes. Finally, I performed ANOVA tests to determine if
macroinvertebrate density and biomass varied predictably by plant species or

architecture.

Results

Using power analysis, I found that by taking an average of 36 samples per
architecture type, I had a power of 1.000 to detect the difference present between the two
plant types (effect size = 0.872). In fact, it would have taken just 7-14 samples within
each architecture type to detect this large difference with a power of 0.85-0.99 (Figure 3).

However, to detect very small differences between the two architecture types (effect sizes

19

.00

der.

Dis.

I.llSI'r

With

mac

= 0.1-0.3) I determined that 60-527 samples were necessary to achieve similar power
(Figure 4a). However, intermediate effect sizes (0.6—0.4) could be reasonably achieved
with 16-34 samples (power = 0.9) (Figure 4a).

For the same analysis of macroinvertebrate density and biomass by plant species,
I found that with our sample protocol (average of 14 replicates per plant species), I had a
power of 0.994 to detect the differences present between the five plant species (effect size
= 0.646). I could have taken just 7-14 samples within each plant species to detect these
differences with a power of 0820-0994 (Figure 3). Similar to the analysis for plant
architecture, 1 determined that 36-310 samples were necessary to detect very small
difi‘erences between plant species (effect sizes = 0.3-0.1) and intermediate effect sizes
(O.6-0.4) could be reasonably achieved with 10-21 samples (Figure 4b).

My results suggest that macroinvertebrate density and biomass are significantly
related to leaf dissection (Figure 5). Dissected-leaf plants (M. spicatum, P. pectinatus,
and Ranunculus sp.) harbored higher densities and biomass of macroinvertebrates than
undissected, broad-leaf plants (P. illinoensis and P. richardsonii) (ANOVA, density P =
.001, biomass p < 0.001). There were no significant differences in macroinvertebrate

density or biomass among plant species within the same architecture type.

Discussion and Conclusion

When planning and conducting ecological experiments, power analysis can lend
insight into how many samples will be necessary to detect differences among treatments
with acceptably high power. I performed these analyses on lentic, epiphytic

macroinvertebrates collected with a mesh bag sampler. I found that I had extremely high

20

power to detect the large differences in macroinvertebrate density and biomass between
the five plant species and two plant architectures (power = 1.000 and 0.994, respectively).
A ‘conservative’ estimate of the number of samples necessary to detect effects would
allow alpha and beta to be set at a level of 0.05, whereas a more ‘liberal’ estimate would
allow alpha to equal 0.05 and beta to equal 0.20 (Peterman 1990b). Choosing an
intermediate of these two (alpha = 0.05, beta < 0.1, resulting in power > 0.9), I
determined that far fewer samples could have been taken within each species or
architecture (9-14), thus allowing time for sampling additional species. I also found that
we could reasonably take sufficient samples to detect intermediate differences among
species or between architectures (10-21 and 16—34 samples, respectively, effect sizes 0.6-
0.4). This knowledge will allow ecologists to better design further studies of epiphytic
macroinvertebrates.

Epiphytic macroinvertebrates and the macrophytes they colonize are ecologically
important components of lake ecosystems. Our results indicate that dissected-leaf plants
harbored higher densities and biomass of macroinvertebrates than undissected, broad-leaf
plants. In Table 1, I summarize some of the past research studying epiphytic
macroinvertebrates in single lakes. These studies sampled fi'om 3-8 plant species, took 2-
85 replicates of each species, and, similar to this study, found that dissected-leaf plants
harbored more macroinvertebrates than other plant types. I had enough information to
calculate power for the study by Gerrish and Bristow (1979). With an alpha of 0.05 and
an N of 10, they had a power of 1.000 to detect the very large differences found between
the three plant species sampled on June 18, 1974 (effect size = 1.87). In fact, the authors

could have detected smaller differences (effect size = 0.5) by taking only 18 samples of

21

each plant species and they could have taken just 3 samples to detect the differences
present (power > 0.9). Knowing this, more time could have been spent sampling
additional plant species rather than replicates within plant species, resulting in more
information about the relationship between macroinvertebrate density and biomass and
plant architecture.

The management of aquatic plants typically involves the removal of plant
biomass either selectively by species or nonselectively. Thus, plant management affects
the abundance and community composition of macrophytes and, consequently, epiphytic
macroinvertebrates. Because these macroinvertebrates are an important source of food
for many species of juvenile fish, an important component of lake food webs, it is
important that ecologists understand the relationship between macrophytes and
macroinvertebrates so that humans may better manage lakes for both plants and fish.
With the knowledge I have gained in this study, I am better prepared to answer questions
such as: Are the patterns seen here between macroinvertebrate density and biomass and
plant architecture common among lakes (Chapter 2)? and, How do macroinvertebrates

respond to changes in macrophyte communities (Chapter 2)?

22

Chapter 2: The interactions between plant architecture and epiphytic

macroinvertebrates and the effects of Eurasian water milfoil invasion

Introduction

The relationship between macroinvertebrates and submerged macrophytes is
partly explained by macrophyte physical structure, also known as architecture.
Submerged macrophytes can be grouped according to architecture based on the number,
morphometry, and arrangement of stems, branches, and leaves (Lillie and Budd 1992).
Macrophytes are diverse in shape and form, and the architecture of plants has been
suggested to influence the colonization of epiphytic macroinvertebrates by providing
macroinvertebrates varying amounts of substrate and cover from predators (Cyr and
Downing 1988a; Jackson 1997). Similarly, macroinvertebrates can be grouped
according to functional feeding groups based on morphological and behavioral
adaptations for food resource acquisition (Menitt and Cummins 1996; Merritt et al.
1996). Although macroinvertebrate functional feeding groups have been used
extensively in lotic systems (e.g. Vannote et al. 1980; Gregg and Rose 1985; Merritt et al.
1996), the use of these groups is much less common in lakes (but see Chilton 1990), and
subsequently much less is known about the relationship between these macroinvertebrate
groups and plants in lakes.

Using functional groups such as macroinvertebrate feeding groups and plant
architecture may help us find and explain patterns between epiphytic macroinvertebrates
and macrophytes. In fact, many studies of plant architecture have found that
macroinvertebrate density is higher on dissected-leaf plants than undissected-leaf plants

(Krecker 1939; Andrews and Hasler 1943; Gerking 1957; Mrachek 1966; Gerrish and

23

Bristow 1979; Cattaneo and Kalff 1980; Dvorak and Best 1982). It has been postulated
that this result is because dissected-leaf plants have a higher surface area to plant weight
ratio and therefore provide more habitat for macroinvertebrate colonization, more food
for grazing macroinvertebrates in the form of periphyton, or additional complexity which
offers better refuge fiom predators (Krull 1970; Pardue 1973; Dvorak and Best 1982;
Gilinsky 1984; Jackson 1997). However, a more recent study found that
macroinvertebrate density did not vary predictably with leaf dissection (Cyr and
Downing 1988a and b). Instead, macroinvertebrate density was plant-species specific
(Cyr and Downing 1988a and b). In addition to surface area to plant weight ratio, other
factors have been suggested to drive macroinvertebrate colonization. In a study of
dissected and thme leafed plants (all with large surface areas per unit weight), Parsons
and Matthews (1995) suggest that the nature of the colonizable surface (soft or hard
stems, brittle or pliable leaves, whorled or non-whorled leaves) is at least as important as
the surface area. Also, a study relating biomass and surface area of six submerged plants
actually found that all dissected-leaf plants did not have higher surface areas than plants
with undissected leaves (Sher-Kan] et al. 1995). Instead, a dissected and an undissected
species (Myriophyllum spicatum L. and Elodea canadensis Michx.) were found to have
higher surface areas than three undissected species (Nitellopsis obtuse Desv.,
Potamogeton lucens L., and Potamogeton perfoliatus L.), and one dissected species
(Potamogeton pectinatus L.). Therefore, the architecture of the plant may not completely
explain differences in surface area (Sher-Kaul et a1. 1995). Although the relationship

between surface area and plant architecture and biomass is not straightforward, because I

24

511

SC.

di:
int

$12

$111
ma
CX(
1'88.

Sm

surface area is difficult to measure directly, surrogates such as biomass or architecture
type are often used instead.

Leaf dissection, or plant architecture, may become important at the whole-lake
scale when the macrophyte community becomes dominated by a single type of plant.
One such plant is Myriophyllum spicatum (hereafter milfoil), an exotic, submerged,
dissected-leaf macrophyte found in much of temperate North America. This exotic was
introduced to North America prior to 1950 from Europe, and by 1985 was reported in 33
states, the District of Columbia, and the Canadian Provinces of British Columbia,
Ontario, and Quebec (Couch and Nelson 1985). Because milfoil has three mechanisms
of propagation (seed production, stolon formation, and fragmentation; Madsen and Smith
1997) and can grow in water from 1 - 10 m deep, it has spread rapidly throughout North
America (Aiken et al. 1979; Couch and Nelson 1985). Milfoil typically forms dense
surface mats or canopies that suppress native plant growth and result in homogeneous
macrophyte beds (Aiken et al. 1979; Madsen at al. 1991). The continued spread of this
exotic species has recreational and ecological ramifications that have prompted much
research into milfoil ecology, biology, and management (Keast 1984; Madsen et al. 1988;
Smith and Barko 1990; Chilton 1990; Trebitz et al. 1993; Madsen and Smith 1997).

Milfoil, a dissected-leaf plant, has a higher surface area for the same unit of
biomass than four other native plant species (undissected: N. obtuse, P. lucens, P.
perfoliatus, and dissected: P. pectinatus; Sher-Kaul et al. 1995). This result would
suggest that milfoil should have high rates of macroinvertebrate colonization. However,
other studies have shown that milfoil actually supports fewer macroinvertebrates than

native plant species (Soszka 1975; Dvorak and Best 1982; Keast 1984; Cattaneo et al.

25

1998). These low macroinvertebrate densities on milfoil may be because milfoil forms
dense homogeneous vegetation beds which, in general, have been shown to support lower
densities of epiphytic macroinvertebrates (Brown et al. 1988). Spatial complexity may
also explain low macroinvertebrate densities associated with milfoil. Dibble et al. ( 1996)
measured frequency of interstices (gaps among stems and leaves) and found that milfoil
had lower spatial complexity than six other plant species (undissected: Egeria densa,
Hydrilla verticillata, Potamogeton nodosus Poir., Vallisneria americana Michx.,
Zosterella dubia L., and dissected: P. pectinatus).

I developed a multi-lake study to answer a few key questions regarding the
relationship between macrophytes and macroinvertebrates, and in particular, how the
spread of the exotic milfoil might affect that relationship. My objectives were: 1) to
document and examine patterns among macroinvertebrates, macroinvertebrate functional
feeding groups, plant species, and plant architectures at a range of scales (within
individual lakes, across six lakes, and for six lakes pooled), 2) to quantitatively synthesize
the published literature on the relationship between macroinvertebrates and plant
architecture, and 3) to examine patterns between macroinvertebrates and the percent
cover of milfoil across six lakes. I hypothesized that l) macroinvertebrate density and
biomass is related to plant architecture, with dissected plants harboring higher
macroinvertebrate densities and biomass than undissected plants, and 2) because dense
homogeneous macrophyte beds may support fewer macroinvertebrates, overall
macroinvertebrate density and biomass will decrease as percent milfoil cover increases

across lakes, even though milfoil is a dissected plant.

26

Materials and Methods
Study Area

Epiphytic macroinvertebrates were sampled from six lakes in southern Michigan
during July and August 1999 (Figure l). The lakes fall along a gradient of percent
milfoil cover (Table 1, see explanations of calculations below). The three lakes low on
the gradient (Camp, Big Crooked, and Lobdell) were treated in May 1997 with 5 - 7 ppb
Sonar® and are part of a study examining the direct and indirect effects of whole-lake
Sonar® treatments on plants, fish, and invertebrates. The three reference lakes were
chosen because they had high percent milfoil cover and have undergone very little plant
management. Two of the three reference lakes (Big Seven and Heron Lakes) are located
in State Recreational and State Park Areas.
Sampling

Macrophytes: Plants were sampled in August of 1999 in the six lakes by the US.
Army Engineer Research and Development Center using the point intercept method
(Madsen 1999). Each lake was mapped using a geographic information system and then
overlaid with a grid of points to be surveyed (150-250 points per lake). Points were
located with a global positioning system. At each survey point, water depth was
measured, a two-sided rake was thrown, and plant species presence/absence was recorded
(J .D. Madsen unpublished data).

Lake Characteristics: To account for inherent differences among lakes, monthly
water quality samples were taken from the deepest area of each lake for nutrients,
chlorophyll, and Secchi depth. The depth of the epilimnion was estimated from

temperature profiles taken each sampling date. A tube sampler was used to take

27

integrated epilimnetic samples for total phosphorus, total nitrogen and chlorophyll a
concentrations. For chlorophyll analysis, water was filtered on site through a glass fiber
filter (W hatman GF-C) and stored in the dark until being returned to the lab and frozen.
Chlorophyll a concentrations were determined fluorometrically with phaeopigment
correction following 24 hour extraction in methanol (Nusch 1980). Total nitrogen was
determined using a persulfate digestion followed by second derivative spectroscopy
(Crumpton et al. 1992). Total phosphorus was determined using a persulfate digestion
(Menzel and Corwin 1965) followed by standard colorimetry (Murphy and Riley 1962).
Secchi disk depth was measured for each lake off of the shady side of the boat.
Macroinvertebrates: Using August 1998 vegetation survey results conducted for
the six lakes (J .D. Madsen unpublished data), the five most common submerged plant
species for each lake were selected for epiphytic invertebrate sampling in summer 1999.
Less common species were sampled in a few cases in order to collect at least two species
within each plant architecture type (dissected and undissected) and to include rrrilfoil in
each lake. I adapted the final list on-site for seasonal and interannual changes that may
have occurred from 1998 to 1999. Each lake was sampled for macroinvertebrates twice
during summer 1999 (June 28 - July 7 and August 16 - August 24). To sample epiphytic
macroinvertebrates, a snorkeller sampled individual plant stems with a 500 um mesh bag
sampler measuring 65 cm long by 24 cm in diameter (Chapter 1). For each lake,
epiphytic macroinvertebrates were sampled at 3 - 5 sites separated by greater than 100 m.
Each site was approximately 2 m deep and consisted of heterogeneous macrophyte beds.
Based on power and sample size analyses from data collected in Heron Lake in August

1998 (Chapter 1), 2 - 4 stems from each of the five macrophyte species were randomly

28

san
resr
Car

Ind

8101

mac
plar
mac
metj
indi

Mac

gram,
a” su]

fleefl

sampled from approximately a 10 m radius around an anchored boat at each site,
resulting in 13 individuals of each plant species, or 65 samples per lake per date (except
Camp Lake in July when only four plant species were sampled), totaling ~800 samples.
Individual samples (macrophyte stem, associated macroinvertebrates, and water) were
stored in a sealed plastic bag and kept cool and dark until further processing.

In the lab, individual macrophyte stems were rinsed with water to detach
macroinvertebrates: the plants were dried at 105 °C for 48 hours and weighed to estimate
plant biomass. Macroinvertebrates were preserved in 95% ethanol. For each lake, the 13
macroinvertebrate replicates from each plant species were pooled and subsampled using
methods developed by Waters (1969). Subsamples were counted until at least 140
individuals had been counted, which resulted in estimates within 20% of the mean.
Macroinvertebrates were identified to the lowest taxonomic level necessary to be placed
into functional feeding groups (scraper, gathering collector, filtering collector, plant
piercer, predator, and shredder; Merritt and Cummins 1996; Merritt et al. 1996). Each
individual was measured to the nearest um using a drawing tube and digitizing tablet.
Macroinvertebrate biomass was estimated from body lengths using length-dry weight
regressions from the literature (Rogers et al. 1977; Smock 1980; Meyer 1989; Burgherr
and Meyer 1997; G.G. Mittelbach unpublished data)

Data Analysis

Macrophytes: Using macrophyte data collected in 1999, I calculated two milfoil
gradients, a non-weighted and a weighted gradient. To develop the gradients, I included
all submerged macrophytes except the macro-a1 ga Chara sp., thus excluding emergent,

free-floating, and floating-leaf plants. For each lake, the littoral zone was defined as the

29

zont
litto
wei‘
had
vegl
not
so i:
grac
of p
entiI
cell.
num
littor
numl
intert
gradi
grid,
milfo
bele
50me\
for the
gr adie

preSer

 

zone from shore to the deepest point at which plants consistently occurred. Within the
littoral zone, Icalculated the percent of sites that were vegetated. To calculate the non-
weighted gradient for each lake, I calculated the percent of the vegetated littoral zone that
had milfoil present (number of points with milfoil present divided by the number of
vegetated sites in the littoral zone multiplied by 100). This non-weighted gradient does
not make any assumptions about the relative density within a grid, only species presence,
so is likely a liberal estimate of percent milfoil cover. I tried to estimate a more realistic
gradient by calculating a weighted milfoil gradient in which I assumed that l) the number
of plant species found at a single point was representative of species composition for the
entire grid cell, and 2) that each plant species was found at equal densities within the grid
cell. Therefore, each milfoil presence count was weighted by the reciprocal of the
number of other species found at that site. I then calculated the percent of the vegetated
littoral zone that had milfoil present (sum of weighted milfoil points divided by the
number of vegetated sites in the littoral zone multiplied by 100). Because the point
intercept sampling method does not assess plant biomass or density, the weighted
gradient was an attempt to approximate the relative milfoil biomass or density within a
grid, and is likely a conservative estimate. The weighted gradient resulted in a lower
milfoil gradient overall; however, the order of the lakes along the gradient did not differ
between the two gradients. The actual percent milfoil cover for each lake is likely
somewhere between these two gradients (non-weighted and weighted). I present results
for the weighted percent milfoil gradient only because regression analyses along both
gradients gave similar results and I felt that, because it considers other plant species

presence, the weighted gradient was more realistic.

30

Macroinvertebrates: For all analyses, macroinvertebrates were standardized by
plant dry weight (g), which allows for the comparison of macroinvertebrates among
different plant species and architecture types. So as to not lose information, I report
results for July and August separately rather than as an average because
macroinvertebrate life cycles are short and periodic, thus density, biomass, and species
composition changes throughout the summer (Gaufin et al. 1956; Mracheck 1966; Soszka
1975; Merritt and Cummins 1996). Macroinvertebrates (expressed as either density (#)
of individuals per gram plant biomass or biomass (mg) per gram plant biomass) were
logm transformed to meet statistical assumptions.

I tested whether macroinvertebrate density and biomass varied predictably by
plant species, plant architecture, and macroinvertebrate functional feeding group at
different scales (within and across the six lakes and pooled lakes) using ANOVA and
adjusted Bonferroni post-hoe comparisons. ‘Within-lake’ analyses refer to
macroinvertebrates in each of the individual six lakes; ‘across-lake’ analyses make
comparisons across the six lakes using the within-lake results; ‘pooled-lake’ analyses
combine data from all six lakes. I also calculated post-hoe power analyses to determine
whether I sampled sufficiently to detect differences in macroinvertebrate density and
biomass among the plant species and architecture types within lakes and with the six
lakes pooled. These post-hoc sample size and power analyses were conducted using
PASS 6.0 software (NCSS Statistical Software 1998). Using one-way ANOVAs and
setting alpha = 0.05, the number of samples that would have been necessary to detect
differences in macroinvertebrate density and biomass between the plant species and

architecture types was estimated at different levels of power and a fixed effect size (the

31

arithn

paran

and l:
macr
macr
epipl

in th

to g
Spec
that
low

Me

rela

kce

ma

gro

0V5

me.

arithmetic difference between the expected value and the observed value for the
parameter of interest).

Regression analyses were performed to determine if macroinvertebrate density
and biomass were related to the percent cover of milfoil. To extrapolate
macroinvertebrate density and biomass to the whole-lake scale, a weighted percent
macrophyte presence was calculated for each plant species from which I sampled
epiphytic macroinvertebrates (the same calculations as for the weighted milfoil gradient
in the macrophyte section above). These numbers were multiplied by the grid cell area
(40-100 m2, depending on lake area) and then by macroinvertebrates per m2 sampler area
to get macroinvertebrate density and biomass per vegetated littoral zone (for that plant
species sampled). For the one case in which I sampled macroinvertebrates from a species
that was not recorded as present at any sites in the plant survey, this plant was given the
lowest experienced occurrence (0.125, or found at one site with eight other species).
Meta-analysis

I used meta-analysis to quantitatively synthesize the published studies on the
relationship between epiphytic macroinvertebrates and plant architecture. In order to
keep personal bias low I did not include selection criteria based on study quality
(Englund et al. 1999). Instead, I included all field studies in which lentic, epiphytic
macroinvertebrates were sampled from submerged plants within the two architecture
groups of dissected and undissected leaves. Although the sampling methods differed
among studies, results from the studies can be compared by calculating a dimensionless
overall effect size by architecture group within each individual study (the ratio of the

means of the two architecture types). Because each study is compared only to itself, the

32

different sampling methods and different approaches should not confound comparisons
(Femandez-Duque and Valeggia 1993; Gurevitch and Hedges 1999).

The published articles I included in these analyses were found using computer
databases: Aquatic Sciences and Fisheries Abstracts (Cambridge Scientific Abstracts),
Biological Abstracts, and ISI Citation Databases (Institute for Scientific Information
Science Citation Index Expanded). Keywords and combinations of keywords included
macroinvertebrate, invertebrate, macrophyte, plant, plant architecture, lake, and lentic.
Older articles not included in these databases were identified and collected from the
reference sections of the more recent articles. Although these extensive searches resulted
in over 75 articles, only 13 articles were included in my final analysis (Table 2). Articles
were eliminated that: 1) did not report/collect quantitative data, 2) did not express data as
numbers of macroinvertebrates per unit of plant biomass, 3) combined benthic and
epiphytic samples, 4) contained data already reported in another article, or 5) aggregated
samples across plant species from different architectures. In addition, of the remaining
13 articles, many did not include all the necessary information (variance) to perform a
weighted meta-analysis, so I performed both weighted and unweighted analyses (see
below). For one of the articles included in the analyses, (Cyr and Downing 1988a) I
obtained raw data from the authors. Because all other studies did not include
zooplankton in analyses, I removed zooplankton from the Cyr and Downing dataset. I
also eliminated samples from their data that aggregated macroinvertebrate samples across
macrophyte species in different architecture groups. Therefore, my conclusions may

differ from Cyr and Downing’s (1988a and b) published conclusions.

33

ob
tin
nu
spt
var
ma

lak

l9}
usit
esti
and
artit
year
data
(Ant

Studi

"1613.
as the

ofllle

Several of the 13 articles included results from more than one lake (Krull 1970;
Cyr and Downing 1988a; Komijow 1989), resulting in 18 lakes as independent
observations (hereafter referred to as studies). For each study, I recorded the number of
times macroinvertebrates were sampled, the decade of sampling, the study location, the
number of species of macrophytes sampled, the number of plants sampled within each
species, the sampler used, the organisms sampled, and the macroinvertebrate density and
variance estimate for each plant architecture group (Table 2). Two studies averaged
macroinvertebrate density across multiple lakes. Because these means included among-
lake variability, meta—analysis was performed with and without these two studies. For
articles that reported data in figure form only (Pip and Stewart 1976; Gerrish and Bristow
1979; Komijow 1989; Chilton 1990), I scanned the graphs and interpolated the values
using Scion Image software (Scion Corporation 1998). For papers that reported multiple
estimates within a season (May-October), densities were averaged across the season (Pip
and Stewart 1976; Gerrish and Bristow 1979; Keast 1984; Chilton 1990). For the two
articles that reported multiple years of data (Soszka 1975; Komijow 1989), only the final
year of data was used to be consistent with all other studies that only had one year of
data. Four studies reported macroinvertebrate density from one sample date only
(Andrews and Hasler 1943; Cyr and Downing 1988a; Chapter 1) and the remaining
studies presented single mean density estimates for the summer season.

I performed two types of meta-analyses: weighted and unweighted. Weighted
meta-analysis incorporates sample variance into the overall effect size (using the variance
as the weighting variable), whereas unweighted meta-analysis does not. Because not all

of the studies included in my meta-analysis provided variance estimates, it was necessary

34

to conduct two meta-analyses: l) unweighted using all 18 published studies and 2)
weighted using the previous seven published studies (except Heron Lake 1998 to be
consistent with my methods of only using a single year of data when multiple years were
reported) and the six individual lakes in this chapter after averaging across months to be
consistent with all other studies. Three of the studies included in the weighted analysis
(published studies only) reported variance estimates directly (Cyr and Downing 1988a;
Chapter 1). For the other studies, I calculated variance by averaging macroinvertebrate
density across two or more plant species within each architecture group (viewing plant
species as replicates within architecture groups as I do in this chapter; Krull 1970;
Andrews and Hasler 1943; Mrachek 1966; Komijowl989).

For both the weighted and unweighted meta-analyses,

effect size = 1 average macroinvertebrate density per plant biomass on dissected plants
average macroinvertebrate density per plant biomass on undissected plants

(Cooper and Hedges 1994; Hedges et al. 1999). The natural log response ratio is
centered around zero (Cooper and Hedges 1994; Hedges et al. 1999). Thus, values
greater than zero indicate that dissected plants have higher densities of
macroinvertebrates than undissected plants.

The weighted meta-analysis was performed with MetaWin (Rosenberg et al.

1997). MetaWin calculates weights for each effect size as (1/ variance) and uses the
weighted effect sizes for hypothesis testing. Because the weighted meta-analysis of 12
studies included estimates of variance calculated different ways, and six studies
performed by one author (Chapter 2), I grouped studies according to author and type of
variance estimate and tested for differences between groups using chi-square tests. Mean

effect size (natural log response ratio) and 95% confidence intervals were calculated for

35

the I
effet
mull
num

mac

Res

Lal

ran
pla
litt.
C01
fro
ran
Cpi
co;

(T;

the unweighted meta-analysis. AN OVA tests were performed to examine whether mean
effect size differed among groups according to the number of lakes sampled (one or
multiple), study area (North America or elsewhere), number of plant species sampled,
number of dates sampled (once, multiple, unknown), organisms sampled (all
macroinvertebrates, snails only, chrionomids only), whether or not milfoil was sampled

(N. American studies only where milfoil is exotic), or decade sampled.

Results

Lake Characteristics

Macrophyte surveys of the six lakes resulted in a non-weighted gradient of
percent milfoil ranging from 21% - 95% and a weighted gradient of percent milfoil
ranging from 4% - 41% of the vegetated littoral zone (Table 1). Each lake had different
plant assemblages, with a range of 10 - 18 submerged plant species in the vegetated
littoral zones and 26 species total across lakes (Figure 2). The percent of the littoral zone
covered with the plant species that were sampled for epiphytic macroinvertebrates ranged
from 56 - 95% (Table l). The six lakes had similar summer mean Secchi disk depth
ranges of 3.3 - 3.7 m, with corresponding photic depth ranges of 8.9 — 10.0 m, and
epilimnion depth ranges of 4.0 - 5.0 m (Table 3). Total nitrogen and total phosphorous
concentrations ranged from 403 - 544 and 15 - 32 rig/L, respectively, in the six lakes

(Table 3).

36

Macroinvertebrates, plant species, and plant architecture

Pooled lakes: In general, the six lakes had similar dominant epiphytic
macroinvertebrate taxa and functional feeding groups (Tables 4 and Figure 3). Thirty-
two total taxa were identified across the six lakes (see Appendix 3), but most taxa were
uncommon (averaging < 1% of total macroinvertebrate density or biomass). After
pooling lakes, 1 found few patterns when I analyzed macroinvertebrate density by
functional feeding groups (ANOVA, Figure 4a and b). In July, I found that no functional
feeding groups were statistically different (Figure 4a, p = 0.15). In August, the density of
predators was significantly higher than gathering collectors, plant piercers, and scrapers
(Figure 4b, p = 0.000, 0.013, 0.039, respectively), and I found a significantly higher
density of shredders than gathering collectors (Figure 4b, p = 0.027). Upon examining
macroinvertebrate biomass, I found that scrapers and shredders generally exhibited
higher biomass than other functional feeding groups (Figure 4c and (1). Specifically, in
July, scraper biomass was significantly greater than predator biomass (Figure 4c,
ANOVA p = 0.007) and in August, scraper biomass was significantly greater than
filtering and gathering collector, plant piercer, and predator biomass (Figure 4d, p = 0.01,
0.028, 0.006, 0.000, respectively) while shredder biomass was significantly greater than
predator biomass (Figure 4d, p = 0.027).

I also examined patterns among the four plant species that were sampled in at
least three of the six lakes. I found no difference in macroinvertebrate density or biomass
(Figure 5, p > 0.273). However, after aggregating all plant species sampled into the two
plant architecture groups (dissected and undissected), I found higher macroinvertebrate

density and biomass on dissected plants than undissected plants (Figure 6 and Table 5

37

last row). Post-hoe power analyses indicated that for analyses that were not statistically
significant at alpha < 0.05, I may have lacked the necessary power to detect differences
between the two plant architectures (Table 5 last row).

Within and across lakes: Unlike the analyses done by pooling lakes, when I
analyzed macroinvertebrate functional feeding groups within individual lakes, I found
that functional feeding groups did not significantly differ (p > 0.1). Similarly, there was
no discernible pattern between average macroinvertebrate density or biomass and
particular plant species within lakes. However, after aggregating all plant species
sampled within single lakes into the two plant architectures, I found patterns of higher
macroinvertebrate density and biomass on dissected compared to undissected-leaf plants
in both July and August, although not all comparisons were significant (Table 5). Post-
hoc power analyses indicated that for analyses that were not statistically significant at
alpha < 0.05, I may have lacked the necessary power to detect the differences between
the two plant architectures (Table 5). In July and August, average macroinvertebrate
density and biomass on dissected versus undissected plants were significantly different
across some of the six lakes (ANOVA; Tables 6 and 7).

Meta-analysis: My meta-analyses corroborate the hypothesis that dissected plants
harbor more macroinvertebrates than undissected plants (Figure 7a). In fact, the
weighted meta-analysis showed that dissected plants have almost twice as many
macroinvertebrates than undissected plants (Figure 7b). I found that studies grouped
according to the number of plant species sampled, the number of times sampling occurred
within a season (once, multiple, unknown), the decade, the organisms sampled (all

macroinvertebrates, chironomids only, snails only), and whether or not milfoil was

38

sampled (N. American studies only where milfoil is exotic) were not significantly
different from one another (p > 0.1 1). In addition, for the weighted meta-analysis, there
were no differences between the two methods of estimating variance (with raw data or
averaging across plant species within architecture groups, p > 0.343) or between studies
conducted by myself versus other authors (p > 0.825). However, the four studies
conducted outside of N. America (Soszka 1975; Dejoux 1983; Komijow 1989) had
significantly lower mean effect sizes than the 14 N. American studies (p = 0.017) and the
two multiple lake studies had significantly lower mean effect sizes than the other 16
studies (p = 0.091) (Krull 1970; Cyr and Downing 1988a).
Macroinvertebrates along a percent milfoil gradient

Along the percent milfoil cover gradient (across the six lakes), I found that the
proportion of dissected plants present significantly increased (Figure 8). If I consider the
relationship between macroinvertebrates and plant architecture alone, then as the
proportion of dissected species increases with increasing percent milfoil cover, I might
expect macroinvertebrates to increase as well. However, when I regressed average
macroinvertebrate biomass and density against the percent milfoil gradient, I actually
found decreasing macroinvertebrates with increasing percent milfoil cover in July, but
not in August (Figure 9). I also examined whole-lake macroinvertebrate density and
biomass using the extrapolated data (whole-lake scale). Although trends were similar
(macroinvertebrate density and biomass decreased with increasing percent milfoil except
for August density), I did not find a significant influence of percent milfoil cover on

macroinvertebrate density and biomass per m2 of the vegetated littoral zone (p > 0.172).

3‘)

Discussion and Conclusion

Using data from my six study lakes, 1 found that lakes had similar dominant
epiphytic macroinvertebrate taxa and functional feeding groups. Scrapers and shredders
generally exhibited higher biomass than other functional feeding groups, especially in
August. I would expect this result because scrapers and shredders use the plant material
and the periphyton that macrophytes provide, and macrophyte senescence in late summer
increases the availability of some of these resources. I also found that higher densities
and biomass of macroinvertebrates were associated with dissected plants than undissected
plants both within lakes and with lakes pooled, although not all relationships were
significant. However, using meta-analysis across a large number of similar studies, I
found very strong patterns that showed that dissected plants had almost twice as many
macroinvertebrates than undissected plants, even when the six lakes from my study were
included. I also found a pattern of decreasing macroinvertebrates with increasing percent
rrrilfoil cover, although the results were not conclusive. Below, I explore the
relationships between plant architecture and macroinvertebrate colonizations, and some
reasons for the equivocal relationships between milfoil and macroinvertebrate patterns.
Macroinvertebrates, plant species, and plant architecture

I hypothesized that plant architecture would be related to macroinvertebrate
density and biomass, with dissected plants harboring higher macroinvertebrate density
and biomass than undissected plants. At multiple scales (within lakes and after pooling
lakes), I found that, although dissected plants exhibited higher densities and biomass of

macroinvertebrates than undissected plants, the patterns were not statistically significant

40

within most individual lakes. Results of post-hoe power analyses indicate that this lack
of significance may have been due to low power.

When designing my study I recognized the high natural variability associated with
macroinvertebrates and tried to maximize statistical power to detect differences between
architecture groups. In each lake, I based my sample sizes on a study conducted in Heron
Lake in 1998 (Chapter 1). A-priori power analyses of Heron Lake 1998
macroinvertebrate data indicated that 9 - 14 samples of each plant species would provide
us with power > 0.9 (alpha = 0.05) to detect differences between plant species and
architecture groups (effect size, the difference between the expected and observed value
of the parameter of interest, > 0.6). To detect smaller differences between plant species
or architectures, I would have needed to take 10 - 21 and 16 - 34 samples per plant
species or architecture, respectively (effect sizes 0.6 - 0.4) (Chapter 1). Because
ecologists do not know what effect sizes are ecologically relevant for littoral zone food
webs, I chose an intermediate 13 replicates per plant species that resulted in statistically
significant differences in Heron Lake in 1998. Contrary to these results, post-hoe power
analyses on the six lakes sampled in 1999 demonstrated that epiphytic macroinvertebrate
density and biomass is extremely variable both across lakes and years. For example, in
Heron Lake 1999, no results were statistically significant at alpha = 0.05, even though in
1998 I had high power to detect differences between the two plant architecture groups in
that lake (Chapter 1). The high variability in Heron Lake may be related to the relatively
large natural decrease in percent milfoil cover Heron Lake experienced between 1997 and

1999 (31%; J .D. Madsen unpublished data). Although the causes are still uncertain,

41

natural milfoil decreases have been documented in other lakes (e. g. Lake Wingra and
southern Ontario lakes; Carpenter 1980; Painter and McCabe 1988; Trebitz et al. 1993).

Although many of the macroinvertebrate/ architecture comparisons in my six
study lakes were not statistically significant (Table 5), the results of my meta-analyses
strongly suggest that dissected plants harbor more macroinvertebrates than undissected
plants. Despite some significant differences among study groups (i.e. studies in N.
American vs. elsewhere), in no analysis did the non-logged 95% confidence interval
overlap zero (which would indicate no difference between dissected and undissected
species) and the weighted analysis indicated that dissected plants harbor almost twice as
many macroinvertebrates as undissected plants. I do not know what is driving patterns of
high macroinvertebrate densities and biomass on dissected plants (e. g. surface area to
volume ratio or structural complexity). However, when I grouped plants into these two
architecture groups, which is much easier to do than measure surface area, Ifound
differences in macroinvertebrate density and biomass. Therefore, using plant architecture
groups as an alternative to measuring surface area may be an appropriate way to
characterize macroinvertebrate density and biomass.

I stated earlier that the four studies conducted outside of North America (Soszka
1975; Dejoux 1983; Kornijow 1989) and the two multiple lake studies (Krull 1970; Cyr
and Downing 1988a) had significantly lower mean effect sizes than the rest of the
studies. Upon removing these six studies, the mean effect size increased from 0.30 (95%
C100] — 0.59) to 0.55 (95% C1022 — 0.89). In contrast, when I calculated the pooled
effect size for the six lakes in this chapter, the mean effect size falls within the weighted

95% confidence interval and supports the hypothesis that dissected plants exhibit higher

42

macroinvertebrate density than undissected plants (Figure 7b). The other two multi-lake
studies may have had low power because there was little replication within architecture
groups and individual lakes. For example, in Cyr and Downing (1988a) only one plant
architecture was sampled in five of the eight lakes. This sample design introduces a high
amount of among-lake variability that may have resulted in low power to detect
differences between the two architecture groups. Although there seems to be no clear
answer for why the four studies performed outside of N. America (Dejoux 1983; Soszka
1975; Komijow 1989) were significantly different from the rest, I offer a few
possibilities: differences in climate (e. g., Africa compared to N. America), differences in
plant species sampled (e. g., P. schweinfurthi is not found in N. America), or differences
in macroinvertebrate populations between regions.

In my study, although I found that dissected plants had higher densities and
biomass of macroinvertebrates than undissected plants, only one lake and two of the four
pooled-lake analyses had statistically significant differences between the two plant
architecture groups. However, when I included these six lakes in the weighted meta-
analysis, I found that all effect sizes were greater than zero and the confidence interval
did not overlap zero. Thus, I conclude that higher macroinvertebrate density is associated
with dissected plants than undissected plants. This result also demonstrates the utility of
meta-analysis to increase power to detect effects and quantitatively synthesize results
across studies.

Macroinvertebrates along a percent milfoil gradient
I found that dissected plants had higher macroinvertebrate densities and biomass

than undissected plants and that the proportion of dissected plants increased across lakes

43

along the percent milfoil gradient. Based on these relationships, I might expect higher
macroinvertebrate densities and biomass in lakes high on the percent milfoil gradient.
However, other studies have suggested that milfoil actually harbors fewer
macroinvertebrates than other dissected plants (Soszka 1975; Dvorak and Best 1982;
Keast 1984; Cattaneo et al. 1989). To ensure that any pattern along the milfoil gradient
was not a result of my sampling scheme, I looked more closely at the proportion of
dissected plants sampled for epiphytic macroinvertebrates in each lake. If the proportion
decreased along the percent milfoil gradient, then my sampling strategy alone might have
influenced the results along the milfoil gradient. However, I found similar proportions of
dissected plants sampled for epiphytic macroinvertebrates across lakes along the percent
milfoil gradient (Table l).

I hypothesized that because dense homogeneous macrophyte beds, such as those
produced in high-milfoil lakes, support fewer macroinvertebrates (Brown et al. 1988),
macroinvertebrate density and biomass should decrease as percent milfoil cover increases
in lakes. My results in July support this hypothesis as macroinvertebrate density and
biomass significantly decreased with increasing percent milfoil cover. However, in
August, these patterns were not evident and after extrapolating my samples to the whole-
lake scale, percent milfoil cover had no significant effect on total lake macroinvertebrate
density and biomass. Epiphytic macroinvertebrates exhibit high natural variability,
which may have contributed to the lack of pattern I found in some cases. For example,
the lake lowest on the milfoil gradient (Camp Lake) was the only'lake that experienced a
decrease in macroinvertebrate densities and biomass from July to August. Upon

examining Camp Lake macroinvertebrates in more detail, I found significantly different

44

insect taxa densities and biomass between the two months (p < 0.000; Figure 9), probably
due to odonate and chironomid emergences between July and August (Figure 10).

Across all lakes in August, these two insect taxa account for approximately 50% and 20%
of the total macroinvertebrate density and biomass, respectively. Although the patterns
were not significant (p = 0.075), the August regressions without Camp Lake were quite
different from those that include Camp Lake and, similar to July, decreased with
increasing percent milfoil cover (Figure 11 versus Figure 9b and (1). Thus, odonate and
chironorrrid emergences may have contributed to my inability to detect a pattern between
macroinvertebrates and percent milfoil cover across the six lakes in August.

To better understand the factors driving the patterns of decreasing
macroinvertebrate density and biomass with increasing percent milfoil cover, Iexamined
macroinvertebrates on milfoil alone. Figure 5 demonstrates that with the six lakes
pooled, on average, milfoil had similar macroinvertebrate densities and biomass as C.
demersum, P. zosteriformis, and P. illinoensis. However, macroinvertebrate density and
biomass on milfoil may be related to the percent cover of milfoil. In general, Ifound that
macroinvertebrate density and biomass on milfoil decreased as percent milfoil cover
increased, although not significantly (Figure 12). This pattern of decreasing
macroinvertebrate density and biomass as percent milfoil cover increased was not
consistent for the other three plant species, suggesting that as milfoil becomes more dense
along the gradient and throughout the summer, colonizable area on milfoil decreases, and
only smaller macroinvertebrates may be able to use the milfoil habitat.

Because juvenile bluegill (Lepomis macrochirus) feed on epiphytic

macroinvertebrates within the vegetated littoral zone (Werner and Hall 1988; Olson et al.

45

1995), I also considered juvenile bluegill densities as a potential driver of
macroinvertebrate density and biomass. If fish densities were driving macroinvertebrate
densities and biomass rather than percent milfoil cover, I would expect to see an increase
in juvenile fish density with increasing percent milfoil cover (and decreasing
macroinvertebrate densities and biomass). However, juvenile bluegill density and
percent milfoil cover were not related (Figure 13; RD. Valley unpublished data), thus
bluegill densities were not the main factor driving macroinvertebrate densities and
biomass.

Because my expectation of fewer macroinvertebrates existing on plants in high
percent milfoil lakes was not consistently proven and could not be explained by bluegill
densities, I considered additional factors that may have confounded the results. For
example, the two lakes lowest on the percent milfoil gradient (Camp and Big Crooked
Lakes) had considerably smaller percent littoral zones and greater mean depths than the
rest of the lakes (Table 3). Although I standardized analyses by the vegetated littoral
zone area when I estimated total macroinvertebrate density and biomass for each lake
(see methods), these morphological differences among lakes may play a role in my
observed patterns. In addition, macrophyte senescence, which starts in late summer for
some plants, may have affected macroinvertebrate density and biomass and confounded
my ability to detect patterns in August. Thus, inherent lake and plant features may have
affected my attempts to relate macroinvertebrates to percent milfoil cover at the whole-
lake scale.

It is important to recognize the limitations of whole-lake ecosystem experimental

designs. For example, Carpenter (1995b) states that for a given experiment, the number

46

of

p6

tw

Sit

he

fa
de
lo
lit
1h.

Pa

of lakes you must sample in order to detect differences depends on the magnitude of the
response factor that is considered significant from biological or management
perspectives. My weighted milfoil gradient ranged from 4 - 41%. From the highest
percent lake to the lowest, this is a difference of only 37%, with the average between any
two lakes along the gradients of only 6%. It is possible that natural inter-lake variation
masked differences caused by such small changes in percent milfoil cover.

Another potential reason I failed to see strong relationships between
macroinvertebrate density and biomass and percent milfoil cover may be found in my
sampling technique. I sampled epiphytic macroinvertebrates from plants in relatively
heterogeneous macrophyte beds. However, milfoil forms dense homogeneous beds
within which macroinvertebrate density, biomass, and taxa richness is higher in the upper
and edge areas than lower and center areas (Sloey et al. 1997). Therefore, I may have
been masking the effects of milfoil by sampling plants from heterogeneous rather than
homogeneous macrophyte beds. Thus, if anything, my results should be conservative and
had I sampled the characteristic dense mats of milfoil, I might have seen stronger
negative relationships between macroinvertebrates and percent milfoil cover.

This study showed that 1) higher macroinvertebrate density and biomass is, in
fact, associated with dissected plants and 2) macroinvertebrate density and biomass may
decrease with increasing percent milfoil. Both of these results have implications for lake
food webs and lake management because macroinvertebrates are an integral component
linking macrophytes, macroinvertebrate-eating fish, and piscivorous fish. Research on
these important food web effects of milfoil on multiple trophic levels and water quality

parameters should improve holistic lake management. However, additional research is

47

needed to further examine the relationship between macroinvertebrates and percent
milfoil cover because my results were equivocal. Based on my findings, Iwould
recommend a study design that includes multiple lakes with larger differences in percent
milfoil, a sampling scheme that includes sampling milfoil from more characteristic dense
homogeneous macrophyte beds, and that include whole-lake macrophyte biomass or

density estimates.

48

CONCLUSIONS AND MANAGEMENT RECOMMENDATIONS

I set out on this research project to address some of the many ecological questions
that have yet to be answered regarding the interactions between macrophytes and
macroinvertebrates. I also wanted to explore how the spread of milfoil, an exotic
macrophyte, and our subsequent management actions to control milfoil might affect these
relationships . My primary conclusions are:

l. The mesh bag sampler I designed to sample macroinvertebrates associated
with submerged plants was relatively easy to use, inexpensive to produce, and
had high statistical power (although power was variable both across lakes and
years; Chapters 1 and 2).

2. Using meta-analysis and data from six lakes, I found that dissected plants
exhibited higher macroinvertebrate density and biomass than undissected
plants (Chapter 2).

3. Macroinvertebrate density and biomass decreased as percent cover of rrrilfoil
increased across six lakes (Chapter 2).

Below I explore how my results can be applied to improve lake management.

The mesh-bag sampler I designed could be used effectively by others to sample
macroinvertebrates associated with submerged plants. This sampler may be a useful
management tool for assessing lake biological integrity as well as a useful scientific tool
for research studies. I hope this sampler will promote further study of epiphytic

macroinvertebrates on submerged plants in lakes.

49

The three lakes low on the milfoil gradient (Camp, Big Crooked, and Lobdell
Lakes), were the result of Sonar® treatments in May 1997. Because I had no control over
management actions taken in these lakes, and each lake was subjected to additional plant
management strategies such as other herbicide applications and mechanical harvesting, I
cannot be certain that differences between reference and treatment lakes were caused by
the Sonar® treatments. Therefore, in chapter 2, I analyzed macroinvertebrate density and
biomass along a gradient of percent milfoil cover. Here, however, I look for differences
between reference and treatment lakes. In July, macroinvertebrate density and biomass
was significantly higher in treatment lakes than in reference lakes (Figure 1a and d, p <
0.002). Similar to the patterns I saw when looking at macroinvertebrates along the
percent milfoil gradient, macroinvertebrate density and biomass in August were not
significantly different between reference and treatment lakes (Figure 1b and d, p >
0.511). Recall that there were large odonate and chironomid emergences in Camp Lake
between July and August (Figures 9 and 10). Similar to analyses in Chapter 2, I removed
Camp Lake from the August data and performed another AN OVA. However,
macroinvertebrate density and biomass remained insignificantly different between

reference and treatment lakes (Figure 1c and f, p > 0.496).

50

July August August excluding Camp

 

 

 

M

A. B. C.

 

 

 

 

 

 

 

Log avg macrornvert
density per g plant bio.
«b
I
I
H—t
l—H

DJ

 

 

 

 

 

W

N
1
o—o—s
l
e—o—t
H—t
l
r—H
r—H

a...
1
0-0—0
T
T

 

 

 

 

 

 

Log avg. macroinvert.
bio. (mg) per g plant bio.

 

 

 

0

Ref. Treat. Ref. Treat. Ref. Treat.
Reference versus treatment lakes

Figure 1. Average macroinvertebrate density (#) and biomass (mg) per g plant biomass in
reference and treatment lakes during July (A and D), August (B and E), and August
excluding Camp Lake (C and F).

Other studies have examined the direct toxic effects and the indirect effects of
Sonar® on benthic macroinvertebrates. In general, direct toxic effects of fluridone on
benthic macroinvertebrates have been negligible. Two studies of Chironomus tentans
larvae found that interactions of fluridone (Sonar® active ingredient) with suspended
solids or sediment had relatively little effect on herbicide accumulation (Muir et al. 1982)
and that fluridone assimilation by these larvae from ingested sediments was negligible
(Muir et al. 1983). Hamelink et al. (1986) found that at the Sonar® label rate (100 ppb)
there was a favorable safety margin between the concentration that affects Gammarus
pseudilimnaeus and Chironomus pulmosus. The only study that found some direct toxic
effects of fluridone on macroinvertebrates (fly larvae; Hydrellia) used fluridone at

concentrations of 4600-9200 ppb. Therefore, although studies have looked only at a few

51

1113'

to .
eff
pre
tha
ma
wa
Slut
add
afie
tre

0011

Ma

efft
Min
Col

ma

 

0ft

Cor

 

 

 

taxa, it appears that there are minimal direct toxic effects of Sonar® on
macroinvertebrates.

Although I cannot attribute the difference between reference and treatment lakes
to Sonar® applications alone, the results of this study indicate potential positive indirect
effects of Sonar® treatments on macroinvertebrate density and biomass. The only
previous study that examined the indirect effects of Sonar® on macroinvertebrates found
that, contrary to the results of this study, fluridone applications resulted in decreased
macroinvertebrate density and diversity (Delong and Mundahl 1996). However, Sonar®
was applied at a much higher concentration of 23 ppb as compared to 5 - 7 ppb in this
study, and there was no replication of treatment lakes ODelong and Mundahl 1996). In
addition, Delong and Mundahl (1996) sampled during the year of treatment and one year
after treatment, during which time there may have been more transient effects of the
treatment present. However, with more replication (three Sonar® lakes) and lower
concentrations of fluridone (5-7 ppb), this study found adequate milfoil control (J .D.
Madsen unpublished data) and potentially positive indirect effects on macroinvertebrates
(Chapter 2).

At the beginning of this study, insufficient data regarding the direct and indirect
effects of Sonar® had not been collected and synthesized. Therefore, Sonar® use in
Michigan had been restricted and debated for nearly a decade. Now, with our
collaborative study nearing conclusion, it appears that Sonar® may be a useful
management tool for lake management of both macrophytes and fish. In fact, at the time
of this writing, the state of Michigan has decided to allow Sonar® use at low

concentrations (56 ppb) as a milfoil management tool (Batterson 2000). Based on the

52

C011

ma

wh

sho

hox

exh

1113

mi]

C01

Cr

 

thr

conflicting conclusions of the two studies examining the indirect effects of Sonar® on
macroinvertebrates (Delong and Mundahl 1996; Chapter 2), it is difficult to conclude
whether Sonar® use has indirect effects (positive or negative) on macroinvertebrates.
Therefore, the indirect effects should continue to be monitored, and management plans
should be adapted as new data are gathered and analyzed.

1 have an additional recommendation for managers faced with decisions such as
how to control macrophytes such as milfoil: because I have shown that dissected plants
exhibit higher macroinvertebrate densities and biomass than undissected plants, managers
may wish to monitor native plant species composition to be sure that after removing
milfoil, dissected plants remain in the lake. Epiphytic macroinvertebrates are an
important forage base for many species of juvenile fish that use macrophyte beds for
cover and as a source for food (Keast 1984; Diehl and Komijowl998; Persson and
Crowder 1998). Thus, a low proportion of dissected plants may cause repercussions

through the lake foodweb from macrophytes to macroinvertebrates to fish.

53

APPENDICES

54

APPENDIX 1

Chapter 1 Tables and Figures

55

 

Table 1. Examples of previous studies examining epiphytic macroinvertebrates
in individual lakes.

 

 

Citation Number of Number of replicates
plant species taken per plant species1

Andrews & Hasler 1943 8 l7
Gerking 1957 3 2
Gerrish & Bristow 1979 3 10
Krecker 1939 7 Variable2
Mrachek 1966 8 Variable’
This study 5 14

 

 

’ Number of replicates taken per plant species for a single sampling period

2 Number of plants sampled not reported, expressed as length of plant sampled
3 Number of samples reported as total for entire summer only (25-85 per plant
species, number of times sampled not specified)

56

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

FIGURE LEGENDS

Epiphytic macroinvertebrate mesh bag sampler that is a modification of the
folding quadrat sampler (Welch 1948). The sampler has the dimensions of 65
X 24 cm and is constructed from 200 um and 500 um mesh, 2 steel rings, and
canvas. It is closed at the bottom by a drawstring.

Five common macrophyte species of Heron Lake, MI, USA. Undissected: a)
P. illinoensis and b) P. richardsonii; Dissected: c) P. pectinatus, d) M.
spicatum, and e) Ranunculus sp. Adapted from Fassett (1957).

The relationship between the number of samples and power at alpha = 0.05 and
effect size = 0.872. The number of samples necessary are indicated by circles
for plant architecture and triangles are for plant species.

The relationship between the number of samples and effect size for a) plant
species and b) plant architecture (alpha = 0.05). Power levels shown are 0.9
(circles), 0.8 (triangles), and 0.5 (squares). For comparison, the lower two
graphs show the enlarged region of 0-50 samples.

Macroinvertebrate a) density and b) biomass by plant species and architecture.
Plant species are abbreviated as: Ill (P. illinoensis), Ric (P. richardsoniz), Pec
(P. pectinatus), Spi (M. spicatum), and Ran (Ranunculus sp.). Bars represent
the standard error for each plant species.

57

L onE
see Ese nausea

A9350 \
3 3638.53 I o
R I 9
@963 .3 Begohsmv . Eu 2

     

wfimeaE ilV
“80:86 Eu em IIIIIV ., .
wet madam

:35 E: com /‘

$9an .3 envenohsmv -
Sausage :8 mm _ umuuflmmmhmu K

wet mmfim \

:88 E: cow

EU m0

 

 

  

 

58

N 2&5

 

.m—
.U
\\
v.
«4w/
W
‘L
a //
\n / .
(
Il\

 

Enema:

 

3.83:..5

59

m ease

3

 

moan—«m me sonaez

Q S w

 

 

36on :85 S

gasses... as: e

 

 

 

mNd

md

mud

 

 

o;

JaMOd

60

a enema

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

main—am me Lona: Z
we ow mm on o on 3 o... mm on R cm 2 3 m o o
Nd
v.0
ed
wd
~
0% o9e 8m com o3 o omm com omm com o2 2: 8 o o
oi oi oi
Nd
m6 n 526m I We n .830Q 16
wd u 830m 4 wd u 830m 4 9o
ad H 530nm 0 ad H wok/Om C wd
o4

 

 

 

 

258:5...“ «ea—.— .m 830% :3..— .<

ms was

61

m ease

258353 can 862: Ear—

 

 

wouoomma wouoommnoeD
0 J W D
am am am am a

 

 

 

 

mas—.8:— .m

sseurorq runld fi/fiur

@808me

 

\

5M

com

/

am

woaoommmveD

 

1

ca

/
a

 

 

+

 

 

cage: 3..

cm

om

ow

om

cm

on

ow

om

sscurorq meld 3/#

62

APPENDIX 2

Chapter 2 Tables and Figures

63

 

 

8 8 3 Ge :. eesem “E .e
9. 9. em 3% mm .329 .m
9. 3. 2 QC 2 use: .4
8 9. em as 2 :25...— .n
coo—e29
2. 9. ms A5 w an .N
9. om 8 :3 e 956 a
“mama‘s b5. 83392352an he.“ Gausumotteeev
45.953 62:83 .5383 3.3:. .3 62969 BEES»
Bea:— vouoommmc e\e 85... 6833:. .x. ones .983: cognac»? .x. .825 :95:— .x. 3:3

 

A enema no Enema: 5:» ueonmohoo 3 68383 0.3 momma .mo>eo_-couoomm% we: 85 moufinoto>5888
omtnaaao com 3388 3an mo «scuba 05 magnum—SE BEES Bea—g vowoommgtueooaon 23. .moumtouomafio nigh—888 8:5 A 033.

64

 

80:03:80 3:003 .&~.~§~\
8.82 6:30.205: 8:058:38

000890828: 30:00 €530.50. Eazémotk: 033
-8002 630:0 00:0 808800 08m £38050: 532838309 808885 8838 0:0 SO

£38050: ESNKQQQELND

83.50838 6:30.205: 3.2055”:
-882 <05 :3 032:2 03 :82 €58.80. 5:38.28 8320 82 8:20

.8. 0.330588%

838032020». :80MQESQR

88035.: =20MS§BK

$20,888.: :80MQESQR
m0889080>8 5:82:03. :80MQESQR ~ .5806
-8002 .<.m.D 42 00:0 wan €02 €58.30. 55:38.23: :80: 502806

8:30.035: :SmmoESQR
8:88:30? :S0MQESQR
.waaexcmsu temmofiusk
623.208: 3:02:30;

 

€380.30. Eazémdtag
3:893:88 882.50% 28088800: mva— 80—33
-8002 .<.m.D 55 00:0 :0: :0BQEN 838030: 532888.00 88:02 0:0 0308:
88:00
.0038: .0055...— .0028:
08088:.» .8 08% 00:5 0. .8385 60—383. 3:03 0:15 8585

 

000380.308 05 :8 0003 00:5
05 3:588 03 8 .090. 0800:2080 0:: >8: mo 860% 82: 000308 00:0: 00:5 05 :05 00:00:08 At 06:80: 0.2. 88000. 0::
855 008800 0:03 88:08: 008: mo 80:88 05 mm 808800 008: mo :38: 05. 00980808 05 8 @3308 006:5 .N 0305

65

 

00005080>8
.8002

000050838
-8002

000050838
-8002

0008:0838
-8002

3::
0280:0030

0000500098
.8002

000050838
-8002

0:0—om

0:03.“

00050
500.5

8800
.805

008850
0005 m

000:00
600000

00800
630:0

2:28

20232

20032

20032

85:88:

00:0

00:0

0880000
0008..:0m

0880.50
0008.:0m

$0: 00003

.000: 0503

08033
.33:

008800 85

008800 08m

5080000 800300000
€500.05. 500830.800
.3003 0000M0E000m

0000000000 00005
.E0E0E00 5530000000
80.00%:0000 200830.890
.00M00N000Q 0000M0£000R

0.3.5000

0000M0E000n~ 0000000000
0.2000500

000000.05. 300030000000

00800000.: 0000m0E000m
0000,2050 0.2000500
€800.05. 80000000000

50.20500
80203800000 rum». 00.002
320%.003000 0000005000n~

i8. 0.203025
60.000350 0000M05000E
0.0003000 0000M0E000m

0000000000 0.205.500 €500.30.
50003002090. 800000000 00003

00000050 000000.500
€800.08. 83300013:

02820

0500005

800880

800880

00:0

00:80:30
000500002

.mow02 :0w004
6:08:00.“ .205 .002
000 .0895 ”000—0— w

mow08083802

$2 3.088:

$2 2.080:

$3 0000M

22 30000
09... 5000

$2 55.0.:

033

88300 0:0 .SU

033

88260 0:0 S0

66

 

08058088
-8002

3:0 0:0:m

08890238

-8002

08058038

.8002

080502008
.8002

m80k..800>8
6:002

08058038
-8002

0:0—om

00050
008802

.<.m.D
£3

.<.m.D .5

.-<..m.D .>Z

.<.m.D $2

.<.m.D S2

202::

20032

20032

2832

20032

2003:

20032

300 0505

003800
005.85

“oz

002.

w0n 0505
0:0 0w005
5:05

mg 08003
0 0002:
:0805

M00 0503
0 008.:
:080—m

00005 0000M05000n~

.00000020m

0000M05000m 50000.03.
5000.00.02.02 800000000 00003

0080200002 0000M05000R
0.000008% 0000M05000R

000000 080M05000a~
55.80% 0000M05000n~
50.20500 50223800209

0000000 0S0M05000m
0000020002 0000M05000n~
50000.08. 500.38.02.02 ”5.20%
00.002 0.000008% 00~0M05000n~
50.20500 500300000200

50000.05. 500382.22
.0.0.§0~\00.\02 00.205 00x02

800000000
0003M 6.2030; 0.203025

0.500 0200000200800
0.080000% 0000M05000n~
.0000..~0\ 0000M05000R
50.20500 5503800200

08.0302

008802

0:0m @0800

0004 .000—0

0:00 000053 .085
:0w00M .005 885
08:08:02 ”000:3 m

0.5.: V0:: 0220

30m 802 0:808:02

3.3 822%

03:
$320 0.0 .00

38 030.0002
0:0 :020m

003 Mon—082

2.2 :99

ova =5

3.3 :5

67

 

 

3 $00 0250

N0: 0 5. 0.0 2. Z 00 G: 00 35 .z 00:. 0.5000 08 .0
B 030

0.8 0.0% 0.: 00. 0.0 00 6.0 00 0.00. .z 00:. 800 520 .m
B 00.00

0.2 08:. 0.0 0.0 E 2 3.0: 0.0 0% .z 5.0. 05000 .55: .0
3 00.00 500520

000 0.5. 0.0 3. 0.0 00 3.5 00 0.02 .z 5.0. 08080 .035 .0
B 8.2 08520

0.2 5.0% 0.0 3. 0.0 R 20.00 2. 0.00 .z 8.9. 53: 00: .0
3 $00

000 00$ 0.5 0... 00 mm a: 2 0.? .z :0. 53.: 0:50 .0

1:01 8 .0300 E 08:: 00
.001 :01 .5 :0 .555: .0050 .0505 a .55: .82 2500.5:
.8. .ze 30%.: -50 £080 .0. 532 8.5 .3500: 8:50 2.0:

.8000 00:82.88 8:03 0053 00 080: 0000000 000 8 0.800 800 0000 05 00 008.000 00 0:00 00.88: 000 0:0 0 0:0—Em :o 8008:: 0:05

0:080:00 8 08008:: 00 00000-0 .:0w80.8 0008 00.0 00:000. ZR 0:0 0:800:80: 008 00.0 00:80 my. 00000808000 0000-0 .0. 800.0-

68

 

 

 

foam $3. $2. 0&2.
000.0005 60200803
0000085 63—2003 ”0200:0003.
008208.... 000200 .220»: 500200 .2200:
0000:0m 602000.005 ”00002080. ”00003902
0000:0m 6023000»: 000005 6029:080E 0000005 68:00ng 000005 60250022
”000000000 000005 60238032 0000085 6020005 ”000008000
00:00:00 0000.05 6020.30 ”0000002004 00:00:00
0002000 .8302 00000080 .9880 0:038 00800.0 0:02:00 0000.02
”0000:0095 00:00:00 080005 60232005 “0000:0095
08000.5 6008280500 3200500 602202 03000.5 6008280500 00:00:00
”0000000 ”0000:0085 ”008000 880806 88000200
0000085 6082003055 0000085 60820200500 0000005 608009300. 0000085 608200—0055
“002880020 ”002880020 ”002805200 “002828020
0mm: 0.... 00% 2....

 

0000803 00050000822008 m5
9.0..» ”8000.. .0:0€0::h .003.

 

32050 8030202200008 .30
0:03 ”8000“ 3:200:5— 600,—.

 

0x00 0003 05 .00 a: 0x08 00000 808800 000 0008 $080200
05 00 00.00: 00380000: .0003 0 :0 000000 0:08 32000.“ 0800003 0000200000 205 0:0 000805 0:0 320000 00050003800008
.3 $00803 000 320000 00030802200008 0800 .00 {con M 300 00:: 0005080828008 808800 0008 0:0 0:3. .0 030,—.

69

 

.2002 0000.000... 000000000000; 030010000000... 60: 00 .080 .0000 028.. 80.0.00

8.00 0 .0000 .020 $.00 0 .200 000.0 0.00 00 .0000 .200 $.00 S .0000 .0000 00.50 05 .0
0.00 00 .2 S .0000 3.00 0 .20... .030 A000 0 .0000 .003 3.00 0 .800 .0000 .320 .0
3.00 0 .0000 .0000 3.00 0 .300 .300 $.00 0 .0000 .0000 00.00 0 .0000 .0: :0 :20: .0
$.00 00000 .0000 .030 00.00 a .0000 .0000 00.00 000 .0000 .0000 00.00 0: .200 .300 .853 .0

3.000.000.0020 $.00000000000 6.000.000.0003 8.000.000.0003 03.8.0 0...: .0

7O

 

 

 

L000 0 .80.: .000... A00 0 .0000 .0: 3 L000 0 .80.: .080 A00 0 .000... .0000 0E5 .0
Z .0038 .0:_0>.m Z .0030: .0:_0>.m 2 00300 620>a 2 Q0300 630>Am 00—00—
:...03. be. .303. 0.2.
000803 0080020>800002 E0800 finger-00,5200:

 

 

 

 

 

 

 

 

 

 

 

5.6m
- 000.0 0000 0000 0000 000.0 - 0000 0000 0000 000.0 0000 000 .0
- - 0:0 0000 0000 000.0 - - 000.0 0000 0000 0000 .860
- - - 0000 000.0 000.0 - - - 000.0 000.0 0000 :20: .0
- - - - 000.0 0000 - - - - 0000 0000 00300.0
0.30020
- - - - - 0000 - - - - - 0000 000 .0
- - - - - - - - - - - .. 93.5 .m
0.95% 63—0900 flu>om EJQEU 00—54
”m .200 I .- w
n 6 9J0? b.0003 a 9.5 000 .36 5.3: 000000 000 0005
0....
93 mom—«0 00 o v SE00 /
088020 .5 85800 BumhmwwfiwMWW> 88202 8005:: 0o 0 200 m 00 a
. as 50330802... 0000 02.000 00: . 00500 as,
0.000 5.0 030 $520 5 0s ”33 0 =0 00 00008000 00 02000000
8
80 00009.0 Ewwfia 0008000000 00003
.050 s .
0000 0000

71

5.6m

 

 

 

 

 

 

 

 

 

- 0000 0000 0000 000.0 0000 - 0000 0000 0000 0000 0000 000 .0

- - 0000 0000 0000 0000 - - 0000 0000 0000 000.0 .3000 .0

- - - 000.0 0000 0000 - - .. 0000 0000 0000 00.3: .0

- - - - 0000 0000 - - - - 0000 0000 0.0200 .0

mums—60:0

- - - - - 0000 - - - - - 0000 0000 .0

- - .. - - - - - - - .. - 95.0 .0

EPWow flow—2WD flo>um 100—020 3101—

.m 6 $033 a 95.0 000 .56 02...: 000200 ”a 0006
0....
GM». nova—WA mO.O V £93 an ago Haw /
U 003 .0 :0 $5505 3m~£oflo>wwOuomMMuoW—meaomvune 3382.0 U020 . ~ 50% do my
050303 00255 m “mama/00 v5 A 35:: 5S» can
0.0.0 5.0 v.2 000 00 03.0 . 0 00:00 00 02000000
0000.0 000 00000000 00000.»
. 00o .
000 000.00 .0 000

72

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

FIGURE LEGENDS

Map of Michigan counties and study lakes. 1 = Camp Lake, 2 = Big Crooked
Lake, 3 = Lobdell Lake, 4 = Heron Lake, 5 = Clear Lake, 6 = Big Seven Lake.

Weighted frequency of plant species in each lake in the vegetated littoral zone
August 1999 (J .D. Madsen unpublished data). Lakes are in order from low
percent milfoil cover (1) to high percent milfoil cover (6). Full scientific
names are: Cabomba caroliana Gray, Ceratophyllum demersum L., Elodea
canadensis Michx., Heteranthera dubia J acq., Myriophyllum spicatum L.,
Najas spp., Potamogeton amplifolius Tuckerm., Potamogeton crispus L.,
Potamogeton foliosus G., Potamogeton gramineus L., Potamogeton illinoensis
Morong., Potamogeton natans L., Potamogeton nodosus Poir., Potamogeton
pectinatus L., Potamogeton praelongus Wulf., Potamogeton pusillus L.,
Potamogeton richardsonii Benn., Potamogeton robbinsii Oakes., Potamogeton
strictz'folius Benn., Potamogeton sp., Potamogeton zosteriformis Femald.,
Ranunculus sp., Utricularia spp., Valisneria americana Michx, Zannichellia
sp. An asterisk (*) indicates plant species that were sampled for epiphytic
macroinvertebrates.

The functional feeding groups by macroinvertebrate density (#) and biomass
(mg) with the six lakes pooled. Percentages refer to the percentage each
functional feeding group makes up of the total macroinvertebrate density or

biomass in July and August, respectively.

Average macroinvertebrate density (#) and biomass (mg) per g plant biomass
within functional feeding groups in July (A and C) and August (B and D).
Data were analyzed by pooling lakes. Bars represent the standard error for each
fimctional feeding group. F. and G. collector stands for filtering and gathering

collectors, respectively.

Average macroinvertebrate density (#) (A and B) and biomass (mg) (C and D)
per g plant biomass for the three (July) and four (August) plant species that
were sampled in 3 - 6 of the lakes. Data were analyzed by pooling lakes. Large
circles represent plant species not sampled. Plant species that are italicized are
dissected and plant species that are not italicized are undissected. Bars
represent the standard error for each plant species.

Average macroinvertebrate density (#) (A and B) and biomass (mg) (C and D)
per g plant biomass for the two plant architecture groups. Data were analyzed
by pooling lakes. Bars represent the standard error for each architecture type.

73

Figure 7. Effect size (natural log response ratio) for each study included in the
unweighted meta-analysis (A) and the weighted meta-analysis (B). An effect
size greater than zero means that dissected plants exhibit higher
macroinvertebrate densities than undissected plants. Filled diamonds represent
the 18 studies included in the unweighted meta-analysis, empty circles
represent the six individual lakes in this study, the filled cross represents the
pooled six lakes fi'om this study (not included in the weighted meta-analysis)
and the filled square is the mean effect size (calculated as the weighted and
unweighted natural log response ratio of the mean macroinvertebrate density
on dissected plants/ mean macroinvertebrate density on undissected plants).
An asterisk (*) indicates those studies that averaged macroinvertebrate density
across multiple lakes. 1 refers to the pooled six lakes in this study, which is
shown for comparison only and was not included in the weighted meta-
analysis. Memph. is short for Memphremagog.

Figure 8. The proportion of dissected plants present across lakes along the weighted
percent milfoil gradient in August (J .D. Madsen unpublished data).

Figure 9. Average macroinvertebrate density (#) and biomass (mg) per g plant biomass
along the weighted percent milfoil gradient in July (A and C) and August (B
and D). Data were analyzed across lakes and each data points represent the
average density or biomass for each plant species within each lake.

Figure 10. Average density (#) and biomass (mg) of odonates and chironomids per g
plant biomass in Camp Lake in July and August.

Figure 11. Total macroinvertebrate density (#) and biomass (mg) per g plant biomass (A
and B) and total macroinvertebrate density (#) and biomass (mg) per m 2
vegetated littoral zone (C and D) across lakes along the weighted percent

milfoil gradient in July and August.

Figure 12. Total macroinvertebrate density (#) and biomass associated with milfoil per g
plant biomass across lakes along the weighted percent milfoil gradient in July

(A and C) and August (B and D).

Figure 13. Cumulative summer (July and August) bluegill densities (#) per m2 across
lakes along the weighted percent milfoil gradient (R.D. Valley unpublished
data). No data was collected fi'om Camp Lake.

74

*l

 

4. Heron
5. Clear

 

 

 

 

 

 

 

 

F l—llllr
OOOOO
(91 gooxofi-N
II—‘
a
r
B
‘0‘.
9" e
E I-
“ U
D on
O .-
"‘ m
4 6i
" a»
” 5
* ‘I'
‘I
. ‘I
rilllll llllll
mOmOmo nomomo
NNv—‘r-t NNv—‘v—d

. 'ds '2
nunquaum 'A
‘dds n

.ds .H

0 - sguuofyarsoz 'd
’— .ds .d

_ snyofyows 'd
” 115‘"th 'd
:ztuosrwqau 21
mnmd ‘J
snSuolamd 'J
mvuyaad 'J
snsopou 'J
suwnu “J
spmougm 'd
snaugulmfi 'J
mayo} '4
.mdsua "1
mzlofltdwv a
'dds 'N
wnamqgs 'W
wmvagds W
”IQ"? 'H
sgsuapvuoa '3
umsuaulap '3
- vunguyoxva '3

 

6. Big Seven

 
 
   
   

a:

 

l
O

100 u
80 i
60 ~
40 —
20 -

.ds .2
vuvoyauw '11
'dds '3

.ds .H
sguuojjxazsoz 'd
.ds .d
WHO/now: ‘J
#9“qu 'd
”WW-’01P“ 'cI
mmm“! 'd
snSuoIamd 'd
snwuyaad 'd
snsopou '(1
sunwu ‘d
sgsuaougm 'J
mugme 'J
9130.110] 'd
sndsya 'd
mnomdwv 'J

_ 'dds 'N
umqugqgs w
ummagds 'W
”W"? 'H
sgsuapnuva '3
wnsxawap '3
vuvguyoxna '3

t

3. Lobdell

 

llll
COCO
«SN—1

501
40*

auoz [mom] pannaBaA u! (parqfigam) Kauanbaag

76

Figure 2

Plant species

Percent of total
macroinvertebrates

Biomass

 

 

Density

 

   

July

Sample Month

     
   
    
   

  

El Plant piercer
I Filtering collector
El Predator

II Scraper

l Shredder

l3 Gathering collector

  

Figure 3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Iii.
0.0 a
.m. u 0 0.50
M II. m
e .m .lol. 0.58695
N wlTl. w TillOii.
m s 008.85
Ma Me i 0.030230 .0
A. ll. A
B D. .IIIOIII. mHOHODSOU .m
3 «f ”1 L0 0 _ _ 0
1.1 1 O 1“. 7.. a... 4.
III.
Till Til. 000002.00
Till. Bonfiom
II.
H Ti m
.M W Tll¢|l|l €00.85
Ill
W “a Till. whOHOO—fioo .mV
J u
o Tiltii J in.
A C E08230 .m
_ _ _ _ 0
0000.05 0... 0 .20 000.000. 000.80.. 0.. 0 .20 00.5 0.050...
. E .93 as 3053.09058an .95 wed
0~6h£0t9>flmahua

Figure 4

Functional feeding groups
78

 

A. July densi

 

 

 

 

6

5
.5

5

5 5
4043 2

B. August density

 

 

 

 

0000003 .020 .0: u .50 £00.00

0020023000000

0: 03.00.00 000

_5
5 . . . .
3.265544332

2

 

 

 

3 5 C. July biomass

. 5
2

5
I 3 o 2

o 1

1

5.
O

 

O

 

 

D. August biomass

 

 

5

i.i

. 3

3

5.
2

2

5
1

max—~53 «aw—a w .3.— AwEv mas-:30.
3020—83203..an Queue?” wed

1

5
0

. 0

Figure 5

Plant Species

79

 

 

 

 

 

 

 

 

TI
01
S
n
e
d
t
m T...
g
u
A
B
Tel.
m.
S
n
e
.0 TI
w.
H
J
A 0 _
5 4 3 2

00083: 0:0—n .w>a w .02. .3056
80.58.0330..an 990.096 wen—

 

 

 

 

 

 

 

 

d
w
c
e
w
0 Ii .0
a n
m U
0
m 0
«.3. Tel. w
m... .m
u D
A
D
d
m
m
$
s m
u
m 0
0 t
al. c
b TI 0
v. .0
.m. D
J
C _ _
3 2 1.. 0

00.353 2:0..— u .02— 8—5 00083:
0.20—0:350..an mug—0.6 as

Architecture

Figure 6

8O

 

 

 

 

 

 

 

 

 

 

I3 I
a:
.2
B-
.22
0e.
:sox
:-
~\
0—1
0‘.
8
En .
en
3
a
D o
O
O
O
O
9
O:
O
O
O
O
fi'Qi-‘I"? 5'2
v—l O O.

 

 

 

 

 

 

 

 

 

 

 

uvapv 3 I “will
.3 II 1'05'331919
:1 0E 11319907
2 6661 “(WH
é. ° 23.33
.5 o
3 P3310043 3.19
d ¢ uanag 31g
3.0104319 5
ouzaas'vgd °\° ouzoasvgd
nqonuvw 3
uoarugdo \
”’ISDIWO
(Clauuvg)
1031;) 117313
mopuapy . mopuaw
866 I “043H
amfvlorm g E
MD g 3
05331018 E if '5 s
, a .. a e
qdwaw a .5 g ¢ 'tldulapv
ugvldumqg E E g ugvldqug
o o
uoogugdo ‘33 ‘3: "‘
e e 3
0933/1’12’ is as 8
b—l t—l o.
)[an Jpnlg II II II
munzaruow O O O 5 .. vumzaruow
0 ‘1'? 3— .8 '0.
v—I o 0'
(ms roam) W1

81

Figure 7

Lakes

 

R2 = 0.890
p = 0.005

 

 

 

7” ,0 <~ .4” .J «J 1“ AU
nu nu nu nu nu nu nu

3:2.— ..Soummmc Ear—omega

15 20 25 3O 35 4O 45

10

0

Weighted % milfoil gradient

Figure 8

82

A. July densi

O - O
.. O O

R2 = 0.236
P = 0 007

 

R2 = 0.014
P = 0 534

.00 r O 0 a

.0 Av.

 

O. O

 

B. August densi

 

 

 

65432T0 6543210

mus—.33 «:2.— w .59 mam—56
0353.359.an «MB—93 wea—

 

0.1
0:00

1

1 m

3

0 0.

__ flu

2

R P

O .0

S
” 00.
0
01
b
w.
u .0.
J

o. O O
C

 

5453525150
4 3 2 1 0

D. August biomass

R2 = 0.046
P = 0.255

O...

. 5

O.

O
‘8

00‘

 

 

. 5

0.0

 

 

5453525150

4

32

1

man—=3: 23—; u .5: 3.5 med—EVE
82:8..33238 mango: we‘—

0

20 30 40 50

10
Weighted % milfoil gradient

0

Figure 9

83

 

 

 

 

 

         

 

 

% 12
a
a g 10 F
g u-
n a
9‘5 3‘ I
5’ E 6 fl JUly
g g0 4 El August
5 e “
a 3‘
. '5 2 ~
3” 5
4 “U 0 i
biomass density biomass density
Chironomidae Odonata
Figure 10

84

Log average macroinvertebrate density (A)

and biomass (B, mg) per g plant biomass

4.5 4

3.5 1

.N
u:

3.5

2.5 “

0.5 q

A. August density excluding Camp Lake

 

 

 

 

 

1.51

 

 

: R2 = 0.054
9 . o p = 0.261 O
O
N 2 8
0 v ‘
O ‘ i
O O
3 O
B. August biomass excluding Camp Lake
9 O
3 R2 = 0.131 0
o
= 0.075 O
\.p\..
’ :
O
. 3
10 20 30 40

Weighted % milfoil gradient

85

 

 

Figure 1 1

 

A. July density

 

O
O

 

R2 = 0.323
p = 0.240

 

 

5
02

2

W51

5.15.0

.l 0

mas—.53 «nu—n a a2. =8=E :c
365—. 82323398.“:— 733 wed

_

B. August density

 

 

 

O

R2= 0.07
p = 0.612

 

 

525.150

2

i“

1

J

O

 

 

C. July biomass

O

O

R2 = 0.534
p = 0.099

 

D. August biomass

 

.5
1

1

0.
3.2.85 :33 w .5; :85: .8 3.5
mas—:3: 8240293335: .53 med

0

 

 

O

O

R2 = 0.408
p = 0.172

 

 

5
1

1

5
0

O

20 30 40 50

Weighted % milfoil gradient

10

Figure 12

86

Bluegill density (log #lmz)

 

 

° 0 R2 = 0.279

P = 0.360

    

 

O

 

5 10 15 20 25 30 35
Weighted % milfoil gradient

87

40 45

Figure 13

APPENDIX 3

Taxa codes, taxa, and functional feeding groups

88

Appendix 3: Taxa codes, taxa, and functional feeding groups.
F. and G. collectors are short for filtering and gathering collectors, respectively.

 

 

Taxa Taxa Functional Feeding
Code Group

11 Insecta, Diptera, Chironomidae, Orthocladiinae Shredder

12 Insecta, Diptera, Chironomidae, Tanypodinae Predator

13 Insecta, Diptera, Chironomidae, Tanytarsini F. collector
14 Insecta, Diptera, Chironomidae, Chironomini G. collector
22 Insecta, Ephemeroptera, Tricorythidae G. collector
23 Insecta, Ephemeroptera, Baetidae G. collector
24 Insecta, Ephemeroptera, Caenidae G. collector
27 Insecta, Ephemeroptera, Heptageniidae G. collector
33 Insecta, Odonata, Coenagrionidae Predator

35 Insecta, Odonata, Libellulidae Predator

40 Insecta, Coleoptera, Curculionidae Shredder
41 Insecta, Coleoptera, Gyrinidae Predator

44 Annelida, Oligochaeta, Haplotaxida, Naididae G. collector
45 Annelida, Oligochaeta, Haplotaxida, Tubificidae G. collector
52 Ostracoda F. collector
54 Gastropoda, Physidae Scraper

55 Gastropoda, Hydrobiidae Scraper

56 Gastropoda, Planorbidae Scraper

58 Insecta, Lepidoptera, Pyralidae Shredder
62 Coelenterata, Hydrazoa Predator

66 Amphipoda, Taltridae, Hyalella Shredder
75 Insecta, Trichoptera, Hydroptilidae, Orthotrichiini Piercer

77 Insecta, Trichoptera, Hydroptilidae, Hydroptilinae Piercer

79 Insecta, Trichoptera, Polycentropodidae, Neureclipsis F. collector
81 Insecta, Trichoptera, Polycentropodidae, Cemotina Predator

82 Insecta, Trichoptera, Polycentropodidae, Paranyctiophylax Predator

83 Insecta, Trichoptera, Leptoceridae, Oecetis Predator

85 Insecta, Trichoptera, Leptoceridae, Leptocerus Shredder
86 Insecta, Trichoptera, Leptoceridae, Nectopsyche Shredder
87 Insecta, Ephemeroptera, Heptageniidae G. collector
88 Hydracarina Predator
89 Turbellaria, Tricladida, Planariidae Predator

 

89

APPENDIX 4

Raw data: macroinvertebrate density

90

 

:d mmd mod. 8.2 2d 2d 8d Rd 2.... 888 .2 88m 23
2d 2d and mi 32 23 a: :4. 8838 .m 88m .23
8.2 de NE 6d 6d 6d Gd 22 N3 9% 888d 6 88: ha
dad Rd 8d Rd 8d 8d 888:: d 86 bi
d3 Ed Ed dad 828 d 86 ea
”2 :2 8.: 9d 8d 3.2 mmd 8:62.28 d 86 23
2d 2d 8d 2d 8d 8.2 8288 .2 86 .23
add :4” 3.8 2d 8d 2: and 2538 6 86 23
3d 8d mm.» as. 8d d: ad ad 888 .> 8.6 23
dmd Rd «2 mad 8d 8d ad 8888 d 686 23
2.3 and 8.2 Ed 23. m2 m2 m: 23. 5388 .2 686 23
2.3 82 3.2 8.2 odd mg 82 8d 888d .6 8.6 23
2d 8d 2d wmd 2d 8d and 8d 6.86.888 2 5% m6 .3:
3d 3d 8d 2.2 Rd add 8d 88:8 d 88 m6 :3
ad :d :d 2d 6d 2d 2d 2d 5388 .2 8.8 m6 >3
6._ dwd 2d 2d 8d dad a; 8888 .m 8.6m 5m 23
8d 3d 2d 2d 2d 3d 2d :d w? 8:888 6 88m ma >3
8d 82 8.: N2 NE 3d 3d 2d «2 88828 d 886 m6 33
Mad 2d 3.2 8d 2d 8d pd Qd 8:838 d 8.86 ma be
8d 2d 2.3 Rd Ed 2d 8d 9d wad «dd 8 .z 8.86 mi 23
8.: 3d 3.6 3d 2: a: Ed 2d 8d 8588 .2 8.86 mi :3
Nwém mm; cNd— th 2d and and cfim mHHH—ohwwm poxoouwxwm anew“
do em mm am mm an an m. N2 __ .

8600 893 com m xmecmmufinww

o a oto>598m€ owao>< 3V .
.83 .558 0388 .3 mdeoE :83 w 8a :3 .3826 a n

and 2a 88% add

91

 

Rd Rd 3.? Rd Rd :dd «.9908 .6 86 892
ddd Rd Ed 2d 2d 2d Rd «8882 99N6 892.
ad ad ddd :d :d 299899.: 956 993..
Rd ddd wddd odd 5: Rd Rd 9288 .2 99...6 892
89 Rd E ddd Rd 996.: 9.86 893
add 3d ed ddd ddd dvd Ed 588d .6 956 892

3d Rd mm: ddd 88888 d 85m 9m unwed
dd: 2: Ed odd «3 ad 8.8989“: 9.5% 9m added
Rd dwd Rd odd d: Rd 9388 .2 85m 9m 892.
3d 3d ddd Rd Rd 3d 9888.”... 88m 9m 89:.
wddd Rd R: 8d a: 9:893 .6 8% 9m :83.
Rd wdd odd wdd odd 3: Rd 8588?: 8986 9m 993
8d 3d ddd 9d dd: Md: 390938 d 8986 9m 9.92
dd: :: wddd R: ddd ddd :: 8.9 99.2 8986 9m 892
:13. 2d 8.: dd: Ed R: odd 2d 9388 .2 8986 9m 9&3.
dddd 8d Rd Rd 3: Rd Rd Rd 9:898 .6 89°86 9m .88
3d 9d: 9d: :d Rd dwd Rd «8882 :83 93
Mad Rd: dwd 2d dmd odd dd: Rd 99.: :88: 93
R: Rd 8.: Rd wad end R: 8.2 :28: 93
R: 8d Ed 8d Ed dd: 9: R: 9388 .2 :28: 93
Rd 2d 3d: 9d 2.9 dwd Rd 9:889 .6 :88: 93
odd 8: odd ddd Ed 9.: 3d manecododd 88: 93
Rd 9d 3d dd: Rd Rd: 3: Ed: 3:89,: :8: 93
8 on mm «d R «d «d 2 d: : 88mm .5: 3:: 5.32

388 9 56:63.9

92

 

Rd ddd odd 2d 2d 2d 2d Rd Rd “80882 :83 893
R: Ed R: 3d Rd 3d 8 .2 :83 893
3: Rd ddd Rd Rd Rd Rd 2899:; :83 893
Rd Rd :3. 3: Rd 2d 3: 8.2 :88: 883
R: Rd Rd Rd Rd 8: R: 888 .2 :80: 893
Rd Rd Rd: 3: R: Rd :3 8888?: 98: 893
RR Rd 3d Rd Rd 8889.: :82 893

ddd RS: odd 8: Rd Rd 988:; 882 893
Rd 3% Rd Rd R: Rd 888 .2 982 :83
RR dad: Rd Rd Rd 8: 888d .6 982 883
3d Rd 2:: Sd 3d S: Rd Rd 8:892 86 8R3
3: :d Rd Rd ddd Rd 988:; 86 893

Rd Rd 8: :d odd 2d Rd mod 9.: 88.883 86 893
R: :d Rd Rd dvd Rd Rd 888 .2 86 893

8 R R R R R R d: d: : 88mm 8: 8...: 5.52

2.286 9 £883

93

 

:d ::d :d Rd 9388 .2 :8: 93

td Ed 2d 98:88 .m 98:: 93

88.58% .U 283.: 2:2.

8: R: dddd Rd 9899:: .5: 86 93
Rd Rd Rdd Rd Rd 98:5: ..: 86 93
R: RR R: Rd 88:38 .n: 86 93

Rd 8.: RR Rd 888 .2 86 93
3: RR Rd Rd 58:98 .6 86 93
Rd Rd Rd: RR R: «888 .> 986 93
2: Rd Rd 2.2 Rd R: Rd 8928:: .5: 986 93
Rd RR Rd Rd 888 .2 986 93

8d Rd dddd Rd 888:6 .6 986 93

Rd Rd Rd Rd 983.88 d 8R 99 93
Rd Rd RR Rd Rd 888.2 d 88 9m 93
Rd :dd Rd: R: 888 .2 89R 9m 93
Rd R: ddd Rd 98:88 .m 88m 9m 93
Rd :dd Rd d:.: 888:V .6 98m 9m 93
Rd Rod :3 :dd :3 8838.8 d 8986 9m 93

Rd Rd :dd Rd 88:28 d 8986 9m 93

RR Rd R: R: .8 .2 8:86 9m 93

Rdd ddd Rd d:d 888 .2 886 9m 93

Rd :3: Rdd Rd 8838 .6 8:86 9m 93
R dd 3. d: R R dd R R dd 88% 8: 3:3 5:32

333 d 9:25:23

94

 

mms 36 find 3.x «Enema .0 .320 “mama/x
ddd Rd ddd d:d ddd 8885 956 and?
ddd ddd ddd ddd dd: ddd :d 88:89,: 986 2.83.

ddd ddd dd: ddd ddd Rd 8388 .2 986 added
odd end mud odd mud 36 «33.3 950 “963‘
ddd add $3 ddd 8:888 .6 986 mama/d
Rd: dddd ddd Rd £8383on 89% 3m .83
ddd dddd ddd ddd: 8889,: :93 d6 added
mmd omd 34m wag: 8533692 co>om wmm “mama/x
ddd Rdd dd: Rdd ddd maaoddsod 83m dB 823
Ed mud who mw.m 8:883 .0 §>om Em «damsdd
dwd cad mi: 23 wmd wad dmd manomuofiod .m 68680 wmm “dams/d
d: d: ddd ddd ddd 33:38 d 886 mi added

Rd ddd: ddd ddd ddd .82 886 m6 added

m5. 3.3 ~m.m end mm: 533m: 68680 Em «damsdd

nmd owé mod 5336866 .0 68380 wmm dm=w=<

dd: Ed R: 2d ddd «8.885 830: 93

and wmdc hmd and dam .D 533 33.

3.: 3.: 3.0m 5d dam .2 $364 bi.
no.3 mod Samoan .2 zovnoq b3.

dwd odd 2 .3 £6 33 add 96 8:80:86 .0 533 b3.
cmd :3 56 mm: £885de .m .863 DE.

ddd dd: 3:8 d 8...: 93

R dd dd d: dd dd dd 2. dd dd 88% =3: 3.3 5.52

353 v 56:39::

95

 

 

ddd ddd ddd: 2: Ed «8885 :88: .83:
ddd: ddd ddd ddd Rdd .8 .6 8:8: .88

ddd Rdd Rddd 2d 8805:: d 833 8d:
ddd Rd dddd d: .8 .z :28: .88
:dd 2.8 ddd :dd :dd 8388 .2 833 88%
Rd: dd: ddd mauodfidad 88 ~83.

hm; maafimuoon .nm €0.53 “ma—waded

ddd ddd 888:: .n: 88d: :83.

ddd ddd ddd 8388 .2 88 ~83:

mmd mwd 885826 .0 modem dd:m=<

dd: dd: dddd d: ddd 3:89,: 86 882
add :dd ddd Rd ddd dd: ddd £882; 86 :83.
odd :dd dd: ddd ddd 8838mm 86 :82

ddd ddd dd:d ddd ddd 8388 .2 86 :88
R dd dd d: dd dd dd dd dd dd 88% :3: 3:: 5.32

886 d 888:.

96

 

 

ddd ddd 888 .2 :82 93

:d ddd 8888 .m :82 93

dd: :dd :dd 888d .6 d8: 93

dd: Rd Rd Rd $8052 d 86 93

ddd de Ed 288 d 86 93

2d d:d R: :d: 83238 d 86 93

Ed ddd ddd: ddd 888 .2 86 93

Rd Rd «838 .6 86 93

R: ddd Rd 80.88 .> 86 93

dd: Rd ddd 8888 .n: 986 93

de ddd: d:: 8388 .2 “H86 93

dd.: ddd ddd 8: 888d .6 :86 93

ddd ddd d: d 8:888 d 88 d5 93

ddd ddd ddd ddd madéoodd 83d mi 93
S: 888 .2 88d d5 93

ddd R: ddd 2888 .m 85d dd: 93
ddd 3d 888d .6 85d dd: 93

Rd Rd :dd 888888 d 8:86 d6 93

Ed 8 Rd 8838 d 886 dd: 93

Ed ddd ddd ddd ddd .82 886 dB 93

Rd Rd dd: Rd 8388 .2 886 dB 93

Rd Rd 888:. .6 886 dB 93

dd dd dd :d dd dd R dd :d R R dd 88% .5: 3.3 5.32

3:03 d 265%?

97

IL

 

2.2 «dd dd." «838 .6 86 .83
:2 Ed d3 add 80.88 .> 956 8d?

ddd ddd ddd :d 8:208 d ““86 added

ddd Rd 888 .2 956 and?

odd dd; 28 .m 86 8d?

dd; d3 ddd ddd 888d .6 986 H83.

dad ddd 8888 d 88d d6 .83.

89 dd; ddd 888mm 83d mi and?

Rd 23 ddd 888 .2 88d dmm 8d?

ddd add ddd $828.6 88d dB and?

Vmé mod 8368660 556m ma dd=w=<

$9 :2 d: d8888 .m 8386 d6 88d
ddd ddd ddd ddd ddd 88:88 d 8686 d6 “8.8
d; :9 2.: odd ddd .82 82868388
dds dds Ed 888 .2 88.6 d5 #88.
dddd 29 Rd 888% .6 8826 da 883
dd; d_d Ed «888 .> 823 93

add .8 .6 833 93

ddd ddd ddd .8 .z 833 93

Rd 888 .2 :83 93

ddd ddd 888% .6 :83 93

Rd 8.2 28828 .m :8: 93

odd 8:8 d 80m 93

dd dd dd dd dd dd R dd 2. R E dd 88% 83 3.3 £52

386 d 883.

98

 

wvd de «8268 .> 5364 “damsad

8.6 86 dad .3 E693 Ramada

and 8625: .m 533 “dams/x

3: 3; dad .Z :63?— adamsadd

Gd mod 882% .2 zucnoq 6363‘

omd wwd duh—£668 .m :66: “dam—2d

3.6 36563 .m 666$ dd=w=<
cod 86 $2625: .m 56$ ddsmsad

$3 3.6 8:638 .2 :66m “dams/d.

mmd mmd mmd mmd 88866 .0 666$ «dams/w

mm.“ 3.6 :3 dsfidda .m 620 “dam—2d

wmd mmd 2.2 _N.m n: £8683: .m 66—0 ddswz<

89 odd 8:838 d 86 H88

wwd Ned mod 838:6 .2 620 6323

mm mm an 3 mm am R an 3. 5 E. d666nm «an: 3:5 5:62

353 v 52669.»

99

APPENDD( 5

Raw data: macroinvertebrate biomass

100

 

8d 8d 3d 2d E2 8d 8d 3d 8d Rd 88888 .2 858 23
8d 8d mod 2d 8d 8d odd 8d ad 8.88880 .m 882 23
8d 8d 2d 5d mod odd 2d 8d 8d 8d wdd Rd 828888U .o 888 23
Ed 2d 2d :d ddd mdd 2d 8888: d 820 23
2d 8d 8d 8d 8d 8828 d 820 23
2d 8d 8d 2d 3d 2d wdd 8d 8202888 .m 820 23
2d 3d 62 «ad ad odd Ed 88828 .2 820 23
8d 8d Rd 32 odd 3d wdd «2 3888 .o .86 23
ad add 8d 2d :2 3d 8: 8d 8d «8888 .> 850 23
2d 2d dmd 8d ddd dwd 8d odd 88288 d 850 23
Sd dam mod Rd N? 82 8d 3d 8d 5d 8888 .2 880 23
$2 8.8 odd Ed ad ad 8d 2d 2d 88.8882U .o 880 23
mod 5d 2d 8d 8d 8d «dd 8d Rd 8888288 .2 8.8m 2m 23
2d 8d odd 8d 8d add 8d 8d N3 88828 .2 8.8m 8m 23
Rd 3d 8d ad add 8d 8d :d 88.8888 .m 8.8m 2m 23
2d 2d 8d 8d 8d 8d 2d 2d 888850 .m 828 2m 23
8d 3d 2d 8d Ndd 8d 8d 8d 8d 3d 8:888 .u 88m 8m 23
ad Rd 2d 82 odd mmd ddd 8d 8d 88 88888888 .2 8286 8m 23
and 8d 2: 8d odd «5 8d 8d Ed 8202888 d 82286 82 23
dvd d: 32 8d odd dim 8d 2d 2d 8d Ndd :d 8% .2 88.86 8m 23
add 82 Sd d3 3d 82 2d 8d 2d dvd 8888m .2 8886 2m 23
odd odd Rd end «Nd 8d dad 8d 8d 8:888 .o 8820 8m 23
8 2 8 cm mm .8 8 mm .8 2 2 2 888m :8: 8.3 8.82

6.26 83 8.2 m 596%? 8m
8de 828% 83a .83 .5808 6883 22 $8803 :63 w 62 38V "”3832 2826238838 omn6>< “m 82.8269?

101

 

2d mod Rd 8.: 8R Rod Rd 8888.0 .86 883.
ad 22 Rd RR Rd 5d 5d 2d 888882 886 883..
8d 3d Ed ed Rd Rd :d 8882882 A288 883..
mod Rd 2d odd 8d 8d 5d Rd 8888m .2 886 88:4.
mod Rd dmd RR 8d 8d 888.: 886 8.83
2d Ed Rd .82 odd 2d 8d ddd 8880880 886 883.
wdd 2d 3d 3d id 888888882 8:8 2m 8888
ddd Ed 3d 8d 3d Rd 8.2 88828 .2 8.8m 2m 88?.
2d 2d Rd 8d wdd 2d ad 8828802 8.8m 8m 883..
Rd dmd 8.2 8d 3d Rd ad 8808808 85% 2m 88.2
Rd Ed 8d Rd Rd mod 888088 .0 85.8 8m 88%
2d Rd Rd Rd 8d Ed ddd 8d 888883882 6880 2m 88.2
dvd 2d Rd 8: Rd 2d Rd 888288“ d 8880 8m 888
R2 22 Rd Rd Rd 2d 8d 3d ed 88.2 8880 2m 8883.
Rd Rd 82 RR :d 8R 8d 8d own 88828 .2 8880 2m 883.
2d 8d RR 8d :d Rd 5d 2d d: 8880880 8880 8m 8833
2d 8d Rd 3d 3d Rod 3d 2d 888885 833 23
Rd Rd 8d ddd Rd odd 8d 8d Ed 88 .3 :88: 23
Rd 2d Rd 8.2 Rd :3 3d 2d 88.2 :88: 23

Ed :d Ed 2d 5d 8d ddd 2d 88828 .2 :88: 23
Rd 8d dvd 2.2 ddd Rd 8d Rd 88808030 2033 23
odd 8d Rd Rd 8d 2d 8d 2d 888828.82 Sam 23

8d 2d 2: Rd 5d ed 8d Rd 8:88.; .888 23
dd 6 8 cm mm 3. an R R R 2 2 88% :8: 3.3 8.82

353 n 8266898.

102

 

3d Ed 8d 3d ddd 8d 8d 5d mdd :d 8880882 833 88:4
odd mod 22 Rd 8d 8d Rd 88.: 833 888
3d ddd 8d mdd Rod 8d Rd 8.8888 .8 2883 883..

2d 2d dmd d3 ddd Rd 2d Rd 8% .z 2033 8:82
Sd Rd 3d ddd «dd Rd 2d 8888... .2 :88: 883..

8d Rd Rd NE. Rd 2d odd 9d 88888882 88...: 883..
odd odd Rd Rd 3d 8d wdd 88888.8 .88: 883.
3d 3d RR dmd 5d mdd :d 8888... .2 8.88 88.2
Rd 3d 5d 8d :d 8808508 888 883.

8d RR Rd 5d ed Rd odd 88888.0 852 883.
2d Rd 8R Rd 2d 5d wdd :d Rd 8888.8 820 883.
odd 82 Rd 8R :2 Rd Rd 88888 d 886 883.
odd Rd 3d 2d 3d 8d 8d Rmd id 8.8888882 .86 883

8d Rd 83 Ed 8d odd add 8d 8888 .2 820 883.
8 8 8 cm mm 3. «m R R 2 2 = 882.8. 88.. 3.3 8.82

388 m £868?

103

 

ddd Rd Rd 3d odd 88828 .2 808: 23

8d 8d ddd Rd 88880 .m 88m 23

Rd 88888 .0 88m 23

Rd 2d Rd odd ddd 88888 d 820 23
:d Rd ddd 2d 8d 8828 d 820 23
:d 82 Rod :d Rd 888288 d 820 23

odd Rd Rd Sd Rd 88828 .2 820 23
2d 2d ddd 2d Rd 8888 .0 820 23
Rd Sd Rd d3 ddd Rd «808:8 .> 880 23
Ed Rd Rd Rd Rd ddd 2d 2d 888888 .8 880 23
mod 8R Rd 2d 2d 8888 .2 8880 23

Rd ddd Rd 2d 3d 88888U .0 880 23

Ed Rd 2d Rd ddd 8888888 .8 88m 8m 23
8d 2d Rd Rd 8888 .2 888 8m 23
Rd Rd Rd Rd Rd 88880 .m 888 8m 23
ad ddd ddd 2d Rd 888880 .m 888 8m 23
8d 2d Rd 8d 8d 8:888 .0 8:8 8m 23
Rd Rd ddd 5d 2d Rdd 88888.8 d 8880 8m 23

3d Rd Rd Rdd Rd 88.888 d 8880 8m 23

Rd Ed 8d 2d 8d 88 .z 8880 8m 23

R2 Rd ddd 3d 3d 88828 .2 8880 8m 23

8d Rd 3d 2d 8:888. .0 82880 2m 23
R. me 3. 3 mm 8 3 2. mm R R 888% :8: 3.3 582

353 n 0268824.

 

104

 

 

dmd ed dmd ddd 8d 8888 .0 8w20 8&5.
Ndd 8d ed 8d Sd ddd «8080885 850 8&3.
8d 8d 3d 5d 8d ddd ddd Sd 38288.8 880 8&3.

8d ddd 3d ddd ddd Sd ddd 8888832 880 883.
8d 8d 8d Nod ddd Ndd 5d 886: 850 8&2
8d 8d 8d ddd Ndd 888083 .0 850 «883
odd :2 mod 8d 8d $88538 .8 8% am 883.
odd 8d :d 8d 8d 888882 825m 80 0883
3d 3d ad 22 8d 880880 .m 82% 8m 0883
8d 2d odd odd ddd ddd 880888 .m 8.3m 8m 8&3.
Nod Ed 8d 8d 8880808 .0 825m 8m 8.83..
8d Ed 3d Rd ddd ddd 5d 888883de 32880 80 8&3
8d 8d 3d 8d Nod Ndd 8:88:38 82080 8m 88%

3d mmd Rd 8d 8d 8% .z 38080 mi and?

3d 32 dd 8d 3d 8888... .2 88.80 8m .883

Nod mad 8d 8:888 .0 82880 80 .883

odd 2d 2d 8d 8d ddd 8508.885 :0ch 23

8d 8d 3d mod 8% .0 $33 23

mod Nod Ed 3d odd 8% .z €23 23
mod Ndd 8d 8888,“ .2 €33 23

8d 3d 82 mod ddd Ndd 5d ddd 838.8% .0 333 23
8.0 cod 8.0 3.0 8.0 $808538 .m :85 b3

cod mod 86 3:62. .m 880m bi.

F mv 3. 3 mm me 3 3. mm NN an «£0on :3..— 8-«4 5.52

383 m 58%?

105

 

 

ad ad ad 2d ddd ddd «508.0885 :88: 8:83
and odd ddd 5d Ndd odd 8% .D 323 3.83

«dd 3d :2 Sd ddd 88882 .8 323 8&3
Ed 8d 8d 8d 8d 8% .z :83: 8883
Nod 92 2d Ndd Nod 5d 8888m .2 333 3:83
Ed 5d Nod ddd $8882.08 .8 880m 883

odd 8d 8d 88882 .8 880m 8:83

8d ddd Nod ddd 8888” .2 880m 8:23

86 mmmnovaado .m 380$ #5ng

odd 8d ddd 83.38% .0 88»: 88:3

3d :d 2d 3d 8d 8288.8 H.020 3:83
Ndd Rd 2d Ed 8d 8d 2d 5d 3885; H8:0 3&3
mod 8d 2d 8d 2d odd 8:828:83 820 0323

m. w... mm. m. a N. s a so 5...: as 8....
mm an an «30on En..— oxaq 552

38.3 m 030533

106

 

888088 .2 880m 21:.

8882088 .m 8823 32.

8888820 .0 8838 ES.

Sd Rd 85082 .8 880 23
8d Rdd 8888 .8 8.80 23
ddd :d 888888 .8 820 23
mod ddd 88888 .2 880 23

380880 .0 8820 32.

3.8 8508.688 .> 8880 33.

omd 88882828 .m 8880 33.

8d 88888... .2 8880 23

Rd 22 88.58% .0 88.80 23

8d 8888388 .8 858 88 23

888088 .2 826m wmm 33.

odd ddd 8.80880 .8 8:58 88 23

8888888880 .m 88>om ME 33.

8888820 .0 8.38m 85 23—.

Rd 8888288 .8 80.880 88 23

odd 8888888 .8 88°80 88 23

5d dmd 88.2 88°80 88 23

«dd Rod 88888 .2 88880 88 23

Rod 8:888 .0 89.8.0 88 23

R 88 G S a. 88 mm «.8 8888 28.8 8.3 552

9808 n 8808838.

107

 

 

cud 8280280 .U 8820 “8:833

ddd ddd «808.88 .> 888 883

86 88883888 .82 8885 88:83»

888088 .2 8880 88:83..

8330 .m 8880 “mama/x

8d 8d 888.88% .0 888 8883
88880888808 .m 88>8m wmm «8:838

8.0 888088 .2 838m wmm 888838

cud 8888888880 .m 828% ma “mama/w

wad 8888208880 .m 896m wmm “8:318

vmd 8888888 .0 826m 85 5:838

R8 8d 88888088 .8 88°80 88 8883
22 odd 5d 888888 .8 88880 88 8383
8: 2d ddd 3d 88m .2 88880 88 8:83

8d 8d 8388... .2 88.80 88 8.83

dwd and 8880880 .0 88880 88 883

odd 885:8 .> 8853 23

88 .3 8033 23

8d 88 .z :38: 23

88888 .2 :88: 23

888088 .0 833 23

8880.88.88.88 .m 888: 33.

b 82:88 .8 888$ 33
N 88 S. B 2. cm mm mm mm 88025 883 8.3 552

3808 n 3088892.

108

 

882.688 .> .3304 2:833

88 .8 833 883

88888 .8 :83 883

88 .z 833 883

cod 888088 .2 $304 8:833

8280:8888 .8 8083.2 8:838

88808:: .82 8883 «8:884

888088 .2 888m 8:833

26 88820880 .m 883» 888883.

5.0 86 8888880 .0 80883 «88:48
Ed 8d 8:88 .8 820 883

who mug cod 8.88085 .8 8820 8:833

88.8 8332888 .8 820 888382

cod .8580QO .2 .820 Hana»

ha 88 :8 5 E. on mm mm MN 833m 88—.— 3_8‘— 5.52

353 n 0388898

109

APPENDIX 6

Raw data: macroinvertebrate length

110

 

a3 «3 m2 EN 3: SN £3 83 3.3 838% .2 85m :3
EN 85 :2 :3 m: mam $3 OS 880850 .m aeom b3
83. 83 33 m? 23 one a: 33 Sm N3 $3 88088 .o 5.53 23
SN 83 :3 33 Ed 83. 388:: .m .86 33
WE on." 82 can 88:3 .m .320 83
N2 2.~ 2.2 8: 85 5+ 83 3:023“ .m .320 33
83 N: a: 86 SN 33 838% .2 H86 83
23 32 33 86 m3 :3 a; «8:88 .o :36 b3
8% £2 23 SN 3o 86 8% 82 «5058“ .> 9:8 b3
2:. 82 83 N3 35 ohm own “8830:. .m 9:8 83
«3 man «3 3a 2.0 £2 83 RN mod 838% .2 85 23
an 3.2 :.~ Gd 9.». m? 3.0 8.3. maggot .o 996 83
«3 mo.” 8: who 91 3.3 NS. 8.3 888528 .2 83m 8m 23
am a; 8.2 a: a: 23 Rd :3 838% .2 5% 8m 83
Rd 82 m2 a: £3 83 Sm 885350 .m 85m 8m 23
N3 2: a; $2 $3 3% own 880850 .m :38 8m 83
am 33 8: 3.0 82 nod 92. 8.». 83 88088 .0 85m 8m 23
EN 82 n: 82 $6 $3 2.3 8.2. 8% @5558 .m 8886 8m b3
33 83 83 9: woo o: o? 83 8:023“ .m 386 8m 83
2.2 23 R; :3 83 :5 SN Ea a; 33 8% g .z 8886 8m 83
SN :1 8.2 22 So 83 81 m3. m3 8382... .2 888.6 8m b3
23 8: 83 2.2 woo $3 $3 mam 8:883 .0 33.85 8m 83
S 8 cm mm um an an a 2 2 2 88% :3: 3.3 582

£68 85 How m 523923 new
88“ can .8609. :83 .83 .558 0383 3 $383 Ema w Hog 3:5 Ewan: Bfinotgfiobaa owfio>< no 5989?

111

 

88.8 8.8 88.8 83 o: 9% 2888 .u .320 883
83 8.8 82 w?“ 808 83 8.8 «80888 .> 886 883
«3 8.8 88.8 83 8.8 8.: 8882888 .8 8:5 883
m: 8.8 $8 3.8 :8 88.8 :8 882% .2 886 883
ohm 8.8.8 82 . v3 2% 226 .88 886 883
22 8.8 3.8 83 a? 81 82: 8883 .u 886 883
5.8 N; mum m: 88883 .8 82,28 88 883

83 88.8 «on 83 83 8.8 882% .2 88m 88 883
8.8 8: 88.8 8.8 Be Ba 288880 .8 88.8 88 883
8.8 8.8 8.8 8.83 8.8 8.8 88880 .8 88m 88 883
8.8 8.8 m3 8? 8.8 888% .o 85m 88 v.83
3.8 33 $8 82 :3 8: 33 858288 .8 8286 88 883
8.8 2.8 2.8 2.8 mom 5. 882888 .8 B286 88 883
2.8 NS 88.8 2.8 88 8.8 3.8 a? 8% .z B286 88 883
82 2.8 3.8 8.2 2.8 23 Ra 8:: 882% .2 8.3885 88 883
8.8 on 8.8 83 83 83 83 R2. 858082. .0 3286 88 883
8.8 $8 :1 3.8 8.8 2.83 3.8 «8883 .> :28: 83
2.8 a: 8.8 one $8 8.8 83 83 8% .8 88283 83
8.8 NS 8.8 8.8 2.». :1. 82m 8% .z 833 83
2.8 3: 3.8 mg 2.8 8.8 a? mom 882% .2 :28: 83
mom m3 83 $8 82 8.8. can 888% .o 833 83
$8 8.8 88.8 $8 8% 2.8. :3 8882.558 .8 8828 83
«2 8.8 3.8 m2 2.8 83 «mm «mm 8:888 .8 8588 83
a 8 cm mm cm 8 mm 3 a 2 z .2888 828 3.3 5.32

3:08 8 8298883

112

 

88.8 88.8 88.8 88.». 8.8 88.8 88.8 3.8 88.... 888828 .> 883 883
88.8 88.8 8.8 88.8 888 88.8 8% .8 833 883
88.8 88.8 88.8 $8 2.» 88.8. 88.» 28822 .8 =283 883
88.8 88.8 88.8 88.8 83. :8 88.8. 8% .z 833 883
2.8 88.8 88.8 88.8 3% 88.8. 88.» 882% .2 88883 883
8.8 82 88.8 88.8 88.8 28. 2.8 282888 .8 8828 8:83
88.». 88.8 88.8 88.8 3.8 2.8 2888: .8 828 883
88.8 8.8 3.8 88.8 2.». 88.8 882% .2 8888 883
88.8 88.8 88.8 88.» 88.8 288888 .m 8228 883

88.8 88.8 88.8 88.8 88.8 88.8 888% .u 8.58 883
88.8 88.8 8.» 2.8. 88.8 88.8 8.8 :8 8288 .8 820 8883
88.8 88.8 2.8. 8.8 8.8 83. 28822 .8 82w 883

88.8 $8 88.8 88.» 88.8 88.8 88.8 88.8 G.» 8:88:88 .8 820 883
88.8 88.8 8.8 88.8 88.8 88.8 8.8 882% .2 820 883

S 8 88 mm 8 88 88 «a 2 2 2 82888 828 3.3 .232

3:08 o 265%?»

113

 

:1 88.8 88.8 88.8 88.8 88.8 882% .2 8828 838

882 88.8 88.8 88.8 88.8 288888 .m 8888 838

88.8 88.8 8888 .o 88: 8888

88.8 88.8 88.8 88.8 88.8 288888 .8 82o 88:
88.8 88.8 88.8 88.8 88.8 28888 .8 820 88.8
88.8 88.8 88.8 :8 8.8 88.8 88288.8 .8 82o 838
88.8 88.8 :88 _8._ 88.8 882% .2 82D .88
8:. 88.8 88.8 88.8 88.8 88888 .0 82o 88:

88.8 88.8 88.8 88.8 88.8 882 8.8888 .> 880 .88
88.8 88.8 88.8 88.8 88.8 88.8 88.8 88.8 888828 .8 886 8.88
83 88.8 88.8 88.8 88.8 8.8 882% .2 8.80 .88
88.8 _8.8 88.8 88.8 88.8 88.8 88888 .0 880 838
88.8 88.8 88.8 88.8 83 28888888 .8 888 8E 838
88.8 88.8 88.8 88.8 882% .2 888 8288 .88
88.8 3.8 88.8 88.8 88.8 288888 .m 888 8E 8.88
88.8 88.8 88.8 83 88.8 288888 .m 888 82m .88
88.8 88.8 2.8 88.8 88.8 .8888 .o 888 8mm 838
88.8 88.8 88.8 88.8 88.8 88.8 88.8 28888888 .8 82880 828 88

88.8 88.8 88.8 88.8 88.8 88.8 8888888 .8 83.86 828 838
88.8 88.8 88.8 88.8 882 88.8 8% .z 882er 28 88

88.8 88.8 88.8 88.8 88.8 88.8 882% .2 8285 8mm 838

88.8 88.8 88.8 88.8 88888 .0 8885 8288 838

8.. 3. 1 88 88 88 88 88 88 88 88 82888 :88 3.3 82.82

3:08 8 888308518

114

 

88.8 88.8 88.8 82 88.8 2888 .0 820 88838

88.8 :8 83. 88.8 88.8 88.8 88888; 880 8883.8
88.8 88.8 83. 88.8 83 2.8 88.: 88.8 88.8 888828 .8 880 .8838
88.8 :8 8.8 88.8 88.8 3.2 88.8 882% .2 880 8838
88.8 88.8 88: E: 88.8 882 88.8 288.: 880 88:88.

88.8 88.8 88.8 88.8 88.8 8888 .0 880 8888.8
:8 88.8 88.8 88.8 88.8 28888.88 .8 888 8mm 883.
88.8 88.8 88.8 81 88.8 882% .2 888 82m 8888
88: 88.8 88.8 8.8.8 88.8 288898 888 828 888.8
88.8 88.8 8:. 88.8 88.8 88.8 28888.8 888 82m 8838
8.8 88.8 88.8 88.8 .8888 .0 888 828 888.8
88.8 88.8 88.8 88.8 88: 88.8 88.8 8888.888 .8 8820 8E .8838
88.8 88.8 88.8 88.8 88.8 88.8 88.8 8888888 .8 8880 888 888.8
88.8 2.8 88.8 88: 88.8 88.8 8% .z 8880 828 8888
88.8 3.8 88.8 88.8 88.8 88.8 882% .2 8820 2: 8838
88.8 88.8 88.8 88.8 8.888 .0 8880 8mm 888..
88.8 88.8 88.8 88: 88.8 88: 88.8 88888 .> 8883 88:
88.8 88.8 88.8 88.8 88.8 8% .D 8883 8E
8.: 88.8 88.8 3: 88.8 8% .z 8883 838
88.8 88.8 88.8 882% .2 8833 838

88.8 88.8 88.8 88.8 88.8 88.8 88.8 88.8 88.88 .0 8253 838
88.8 88.8 88.8 3: 88.8 88.8 28888.88 .8 88: 88.8

88.8 88.8 88.8 82888 .8 88: 838

8.. 3 E 88 88 8.. 88 88 88 88 88 .2888 88.8 3.3 582

3:08 8 88:88:85

115

 

88.8 88.8 88.8 88.8 88.8 88.8 888885 8883 888.8
88.8 88.8 88.8 88.8 88.8 88.8 88.8 8% .D :88: 883.

88.8 2.8 88.8 88.8 288888 .8 888.8 8838
_8.8 88.8 88.8 882 88.8 8% .z 8883 8883.
88.8 88.... 88.8 :2 88.8 882% .2 :88: 8888.8

88.8 88.: 88.8 88.8 88.8 2888888 .8 88: .888.

88.8 88.8 88.8 88.8 28888: .8 8888 88838

8.8.8 88._ 88.8 88: 88.8 882% .2 8888 888..

SN 88:03:80 .m :08om 888E858

88.8 88.8 :1 88.8 8888 .0 88: 8888.8.

88.8 88.8 83 88.8 88.8 828888 .8 820 882
88.8 88.8 88.8 82 88.8 88.8 88.8 88.8 288888 .8 820 8888.8
88.8 88.8 88.8 88.8 88.8 88.8 882888 .8 820 888.8
88.8 88.8 88.8 88.: 88.8 88.8 882% .2 820 888.8
88 3. E 88 88 8.. 88 88 88 88 88 82888 :88 3.3 582

3:08 8 88:05:98..

116

 

8:380:88 .2 :08: b2.

88:03:80 .m :033 b3.

8:80:88 .0 :88.“ $3.

8.8 88.8 88.8 28888 .8 820 88
88.8 88.8 88.8 2882 .8 820 88:
88.8 88.8 888888 .8 820 8E

88.8 88.8 83 882% .2 820 8.88
88.8 28888 .0 820 88:
88.8 88. : 88888 .> 880 8.88
mod 3 . 8 98:38:: .88 9:80 32.
88.8 882% .2 8.80 838

88.8 88.8 88888 .0 8.80 88:

88.8 88 2888.888 .8 888 82: 88:
mod 838088 .2 55% wmm 33.
88.8 88.8 88.8 88.8 288888 .m 888 8E 838
88.8 88.8 288888 .m 888 82: .88
88.: 8888 .0 888 8mm 838
88.8 28888288 .8 8880 8288 .88

88.8 88888888 .8 88880 8288 838

88.8 88.8 8% .z 8880 828 83

88.8 88.8 882% .2 88820 888 88

88.8 88.8 88.88 .0 8880 8mm .88
88 88 88 S 88 88 88 88 88 88 82888 2.8.8 83 8.82

3:08 8 0308808588

117

 

88.8 88.8 8888 .0 820 8838
88.8 88.8 88.8 8.88885 880 888.8
mg. 3:282: .m 9:80 88:838.

cod 8:380:88 .2 @880 888838
88.8 288 .m 880 888.8
88.: 88.8 88.8 8888 .0 880 888.
mud 888808888808 .m 55% wmm 88:388.
88.: 3.8 882% .2 888 8288 888.8
ofim Nw.m 86:03:80 .m :0>0m Em 88:918.
mm.~ $6 88:03:89.,8 :0>0m Em 88838
VGN mad 8:80:80: .0 :0>0m wmm 88838
88.8 88.8 88.: 2888888 .8 8880 8mm 888
888 8.8 88.8 88288.8 .8 8880 8mm 8838

88.8 88.8 88.8 88.8 8% .z 8880 8288 888.8

88.8 88.8 882% .2 8880 8E 888

88.8 88.8 8888 .0 8880 28 883.

88.8 88288 .> 8883 8E

8% .88 888.8 8E

88.8 88.8 8% .z 8883 88.8
882% .2 8883 .88

88.8 8888 .0 :88: 838
8888208288808 .88 :080m b3.

82—88:: .m :053 23—.

88 88 88 G 88 88 88 88 88 88 82888 .8: 88.8 582

3:08 0 8888:2588

118

 

88.8 888588 .> 353 8883..
8% .D 88035 88:83.

3.8 8888.88: .8 8333 8883..
88.8 888 .z 8833 8888.8
88.8 82 838888 .2 8033 8883.
88888888808 .m no.5! 8889188

88888025: .A 858$ 885w=<

838298 .2 88.53 8338

min 86:38:80 .m 888$ «mama/8

88.8 88.8 5:83.88 .0 8808 88:88.8

88.8 88.8 88.8 8:88.88 .8 8820 888%
88.8 8.8 88.8 8.8 8888855 .8 886 .3838
88.8 88088888 .8 836 83838

88.8 88.8 838888 .2 886 8:83.
88 88 88 8.. 88 88 88 88 88 88 88888 2.8.8 8:3 5.82

3:08 e 886585488

119

APPENDIX 7

Raw data: macrophytes

120

 

121

88.88 88.88 88.8 88.8 88 2.88.888 2 88888 838
88.8 88.88 88.8 88.8 88 88888888 .8 88888 8888
88.8 88.88 88.8 88.8 88 5888588 U 88888 88.8
88.8 88.88 88.8 88.8 88 8.888888 .8 8886 8888
88.8 88.88 88.8 88.8 88 8.88.888 .8 886 838
88.88 88.88 88.8 88.8 88 888888858 .8 886 8888
88.8 88.88 88.8 88.8 88 58888888 .2 886 8888
88.8 88.88 88.8 88.8 88 8.3.828 U 8886 8888
88.8 88.88 88.8 88.8 88 888.558 .8 8880 8888
2.8 88.88 88.8 88.8 88 888888888 .8 8880 8888
88.8 88.88 88.8 88.8 88 558.888 2 8&6 8888
88.88 88.88 88.8 88.8 88 3828588 U 888 838
88.8 88.88 88.8 88.8 88 8882882288 .8 88886 888 838
88.8 88.88 88.8 88.8 8_ 38888858 .8 88886 888 8888
88.8 88.88 88.8 88.8 88 .888 .2 88886 888 8888
88.8 88.88 88.8 88.8 88 8888.888 .2 88886 888 838
88.8 88.88 88.8 88.8 88 2828588 .8 888880 888 838
88.8 88.88 88.8 88.8 88 88:88:28.8 .8 8258 888 8888
88.8 88.88 88.8 88.8 88 388888 .8 8858 888 838
88.88 88.88 88.8 88.8 88 5.88.888 2 8268 888 838
88.8 88.88 88.8 88.8 88 88888888 .8 8288 888 8888
88.8 88.88 88.8 88.8 88 2828288 .8 88888 888 8888
mafia—GE Each. .>.U Gd amass—omn— anE Z 8309» “Ea—8— uu—aw— £232

828% 8983 E88 .83 .6888 83 888805 88388888 :38 88:8 .A.>.Uv 858888888, 80 2886588 Addv
8288386 8:888:58 .va 888803 2838808886 :82: .2 :83 8288888 80 8858:: 2: .88“. 883880.882 ”8 88658818

 

88.8 :88 2.8 88.8 88 8.3.8888 U 8886 888.
88.8 88.88 88.8 88.8 88 888888888 .8 8&6 838
88.8 88.88 88.8 88.8 88 8888888888 .8 8&8 888.
88.8 88.88 88.8 88.8 88 8588.888 2 8&8 838
88.8 88.88 88.8 88.8 88 8.88% .88 8&6 88.8
88.88 88.88 88.8 88.8 88 28.22888 .9 8580 888.
88.8 88.88 88.8 88.8 88 88588888288 .8 88886 888 88.8
88.88 88.88 88.8 88.8 88 888888858 .8 888886 888 888.
88.8 88.88 88.8 88.8 88 .888 2 88886 888 888.
88.8 88.888 88.8 88.8 88 558.888 2 8888.6 888 838
88.88 88.88 88.8 88.8 88 5888588 0 88885 8E 88.8
88.8 88.88 88.8 88.8 88 8858882288 .8 82,88 888 888.
88.8 88.88 88.8 88.8 88 388.888 .8 8268 888 838
88.8 88.88 88.8 88.8 88 8588.888 .2 8268 888 838
88.8 88.88 88.8 88.8 88 8.588888 .8 88888 888 88.8
88.8 88.88 88.8 88.8 88 5888588 U 8888 888 838
88.8 88.88 88.8 88.8 88 888.528 .8 8838.8 288
88.8 88.88 :8 88.8 88 .888 .2 88883 8888
88.8 88.88 88.8 88.8 88 .888 2 88888.8 8888
88.8 88.88 88.8 88.8 88 358.888 2 88885 8888
88.8 88.88 88.8 88.8 88 58888588 .8 88883 8888
88.8 88.88 88.8 88.8 88 8858885888 .8 88888 8888
88.8 88.88 88.8 88.8 88 388888 .8 88888 8888
mus—~33 36H. .>.U Gd anus—33 8302 z «39on «Ea—8— Ou—a‘u £232

£883 8 5288.98

122

 

88.8 88.888 88.8 88.8 88 888.588 .8 8828.8 838
88.8 88.88 88.8 88.8 88 .888 .8 88883 82
88.8 88.88 88.8 88.8 88 88888888 .8 88883 888~
88.8 88.88 88.8 88.8 88 .888 .2 8888.8 884
88.8 88.88 88.8 88.8 88 8588.888 2 88883 88.8
88.8 88.88 88.8 88.8 88 8882885888 .8 88888 838
88.8 88.88 88.8 88.8 88 388.888 .8 88888 838
88.8 88.88 88.8 88.8 8_ 88288.58 .8 88888 88<
88.8 88.88 88.8 88.8 88 8588.888 2 88888 83
88.88 88.88 88.8 88.8 88 58888888 .0 88888 88<
88.8 828 88.8 88.8 88 38:88 .8 8886 882
88.8 88.88 88.8 88.8 88 8.588888 .8 888 88.8
88.88 88.88 88.8 88.8 88 888888828 .8 886 88<
88.8 88.88 88.8 88.8 88 8588.888 .2 886 838
mun—EOE 13°F .>.U G.m gun-OE nae—Z Z 839on «flu—8— 3:81— 5:92

35$ 8 £26898

123

APPENDIX 8

Aggregated raw data: macroinvertebrates per plant stem

124

 

88 88 88 88 88.88 88.8 8888: 888.5888 8
88 88 88 88 88.88 88.8 85 888 .2
88 8 88 88 88.88 88.8 888.5 88888888288 .8
88 8 88 88 88.88 88.8 8885 882888 .8
88 88 88 88 88.88 88.8 8885 8888888888 .8
88 : 8 88 88.88 88.8 85 388888 .8
88 : 88 88 88.88 88.8 88885 88888.88 .8
88 88 88 88 88.88 88.8 888.5 8.88.888 .8
88 8_ 88 88 88.88 88.8 888.5 88882858 .8
88 : 88 88 88.88 88.8 8888: 888 .2
88 88 88 88 88.88 88.8 85 8588.888 2
88 8 88 88 88.8 88.8 88885 8.8888 .8
88 88 88 88 88.8 88.8 88885 8.838838 .8
88 8 8 88 88.88 88.: 88 88.28588 .8
88 8 88 88 88.88 88.8 85 8.3.828 .8

.3ng sown—gum—O staunch—many .3008:— bmmflofi mans—SE no.3 @3927. «Ea—8—

.x. 8X. 8X. .X. 4508,5982": 989258:an wan—m

.883 083 9:8 883 8888 05 88:88 82888 3883280 0.63
muwflaooba 88:8 80w88o>< 8088888 8:83 :80 no 88398» $838008 1888:2388 85m 05 .80 £880 a: 8.88 885 85228

05 88:8 8888:0033 98.8 .mvoaobmaw .3885 28 :85 8885283508888 0805 .80 8888828 85 6288888 828888 some 80.8

8583 mam—888 8 no 988.8 9 8898 :8me so» 8.5 888803 98 95 888888828 8858826808088 ow880>8 2:. ”w 8888:2548

125

32688888 £90238 wgzm Ea wstofimw 8.8 88028 88 888:8 .m can .0 ._.

 

88 8 88 88 88 88 28.5 8:88.558 .8
88 88 _8 8 8_ 88 85 888 .88
88 88 88 8 88 88 8825 85388288 .8
88 88 88 : 8_ 8_ 88:5 38:88 .8
88 8 88 8 88 88 8888: 38888888 .8
88 8 88 8 88 8 85 388.888 .8
88 8 88 8 88 88 888.5 8.388% .8
88 8 88 88 88 88 8885 888888 .8
88 8_ 88 8 8_ 8_ 285 888888828 .8
88 8 88 8 88 88 £85 888 .2
88 8 88 8 8_ 88 85 8588.888 .88
88 8 88 8 88 88 888.5 8.5% .8
88 8 88 8 88 88 888.5 8.588888 .8
88 8_ 88 8 8_ 88 85 58888588 .8
88 8 88 8_ 88 88 85 8.38288 U

heuafloha «Leno—=80 .u «havoc—=5 .u hoohomm hot—89:: LOAF-um :9; was—.— 83on ans—8m

89359.3 95..» «5638 3:33:58

 

9:03 w 5988888

126

LITERATURE CITED

127

LITERATURE CITED

Aiken, S.G., P.R. Newroth, and I. Wile. 1979. The biology of Canadian weeds.
34.Myri0phyllum spicatum L. Can J Plant Sci 59: 201-215.

Andrews, J. D. and AD. Hasler. 1943. Fluctuations in the animal populations of the
littoral zone in Lake Mendota. Transactions of the Wisconsin Academy of
Science, Arts, and Letters 35: 175-185.

Barko, J. W. and W. F. James. 1998. Effects of submerged aquatic macrophytes on
nutrient dynamics, sedimentation, and resuspension. The structuring role of
submerged macrophytes in lakes. J eppeson, E., M. Sondergard, and K.
Christoferson Eds. New York, Springer-Verlag: 197-213 (423 total).

Batterson, T.R. February 2000. An update on Sonar®. Michigan Riparian.

Beckett, D.C., T.P. Aartila and AC. Miller. 1992. Seasonal change in plant-dwelling
Chironomidae and Naididae in a Wisconsin lake. Journal of Freshwater Ecology
7(1): 45-57.

Biggs, B. J. F., and T. J. Malthus. 1982. Macroinvertebrates associated with various
aquatic macrophytes in the backwaters and lakes of the upper Clutha Valley, New
Zealand. New Zealand J. of Marine and Freshwater Research 16: 81-88.

Brown, C. A., P. Thomas, J. Poe, R. P. French HI, and D.W. Schloesser. 1988.
Relationships of phytomacrofauna to surface area in naturally occurring
macrophyte stands. J. N. Am.

Benthol. Soc. 7(2): 129-139.

Burgherr, P. and Meyer El. 1997. Regression analysis of linear body dimensions vs. dry
mass in stream macroinvertebrates. Archiv fur Hydrobiologie 139(1): 101-112.

Carpenter, SR. 1980. The decline of Myriophyllum spicatum in a eutrophic Wisconsin
Lake. Can. J. Bot. 58(5): 527-535.

Carpenter, S.R., Ed. 1988. Complex interactions in lake communities. Springer-Verlag,
New York, NY. 283 p.

Carpenter, SR. 1989. Replication and treatment strength in whole-lake experiments.
Ecology 70(2): 453-463.

Carpenter, SR. 1996. Microcosrn experiments have limited relevance for community
and ecosystem ecology. Ecology 77(3): 677-680.

128

Carpenter, SR. and MS. Adams. 1979. Effects of nutrients and temperature on the
decomposition of Myriophyllum spicatum L. in a hardwater eutrophic lake.
Limnol. Oceanogr. 24: 520-528.

Carpenter, SR. and D.M. Lodge, 1986. Effects of submersed macrophytes on ecosystem
Processes. Aquatic Botany 26:341-370.

Carpenter, S.R., S.W. Chisholm, C.J. Krebs, D.W. Schindler and RF. Wright. 1995b.
Ecosystem Experiments. Science 269: 324-327.

Carpenter, S.R., P. Cunningham, S. Gafny, A. Muoz-Del-Rio, N. Nibbelink, M. Olson, T.
Pellet, C. Storlie and A. Trebitz. 1995a. Responses of bluegill to habitat
manipulations: power to detect effects. N.A. Journal of Fisheries Management
15(3): 519-527.

Cattaneo, A. and J. Kalff. 1980. The relative contribution of aquatic macrophytes and
their epiphytes to the production of macrophyte beds. Limnol. Oceanogr. 25(2):
280-289.

Cattaneo, A., G. Galanti, S. Gentineta and S. Romo. 1998. Epiphytic algae and
macroinvertebrates on submerged and floating-leaved macrophytes in an Italian
lake. Freshwater Biology 39: 725-740.

Chilton II, E.W. 1990. Macroinvertebrate communities associated with three aquatic
macrophytes (Ceratophyllum demersum, Myriophyllum spicatum, and Vallisneria
americana) in Lake Onalaska, Wisconsin. J. Freshwater Ecology 5(4): 455-466.

Chick, J. H. and C. C. McIvor. 1994. Patterns in the abundance and composition of fishes
among beds of different macrophytes: viewing a littoral zone as a landscape. Can.
J. Aquat. Sci. 51: 2873-2882.

Cohen, J ., 1988. Statistical power analysis for the behavioral sciences. Erlbaum
Associates, Hillsdale. 415 pp.

Cooke, G.D., A.B. Martin, R.E. Carlson. 1990. The effect of harvesting on macrophyte
regrowth and water quality in LaDue Reservoir, Ohio. J. Iowa Acad. Sci. 97(4):
127-132.

Cooper, H. and L.V. Hedges Eds. 1994. The handbook of research synthesis. Russell
Sage Foundation, New York 57 3 pp.

Couch, R. and E. Nelson. 1985. Myriophyllum spicatum in North America. In: L.W.J.
Anderson (ed), Proceedings of the First International Symposium on the
Watermilfoil (Myriophyllum spicatum) and related Haloragaceae species. Aquat.
Plant Manage. Soc., Washington, DC. pp. 8-18.

129

Creed, R.P., Jr. and SP. Sheldon. 1992. The effect of the weevil Euhrychiopsis lecontei
on Eurasian watermilfoil: Results from Brownington Pond and North Brook
Pond. Proceedings, 27h Annual Meeting, Aquatic Plant Control Research
Program, Miscellaneous Paper A-93-2.

Crowder, LB. and WE. Cooper. 1982. Habitat structural complexity and the interaction
between bluegills and their prey. Ecology 63(6): 1802-1813.

Crumpton, W. G., T. M. Isenhart, and P. D. Mitchell. 1992. Nitrate and organic N
analyses with second-derivative spectroscopy. Limnology and Oceanography 37:
907-913.

Cyr, H. and J .A. Downing. 1988a. The abundance of phytophilous invertebrates on
different species of submerged macrophytes. Freshwater Biology 20: 365-3 74.

Cyr, H. and J .A. Downing. 1988b. Empirical relationships of phytomacrofaunal
abundance to plant biomass and macrophyte bed characteristics. Can. J. Fish.
Aquat. Sci. 45(6): 976-984.

Dale, HM. and T.J. Gillespie. 1977. The influence of submersed aquatic plants on
temperature gradients in shallow water bodies. Can. J. Bot. 55: 2216-2225.

Dejoux, C. 1983. The fauna associated with the aquatic vegetation. pp 273-291 In
Carmouze, J .P., J .R. Durand, and C. Leveque Eds. Lake Chad: Ecology and
productivity of a shallow tropical ecosystem. Dr. W. Junk Publishers, The
Haugue/Boston/Lancaster 575 pp.

Delong, MD. and ND. Mundahl 1996. Secondary Effects of fluridone treatment on
invertebrate community structure in lake ecosystems. Minnesota DNR Technical
Report

Dibble, E. D., K.J. Killgore and GO. Dick. 1996. Measurement of plant architecture in
seven aquatic plants. J. Freshwater Ecology 1 1(3): 31 1-318.

Dibble, E. D. and S. L. Harrel. 1997. Largemouth bass diets in two aquatic plant
communities. J. Aquat. Plant Mgrnt. 35: 74-78.

Diehl, S. and R. Komijow. 1998. Influence of submerged macrophytes on trophic
interactions among fish and macroinvertebrates. In J eppesen, E., M. Sondergard,
and K. Christofferson (eds), The Structuring Role of Submersed Macrophytes in
Lakes. Springer-Verlag, New York: 24-46.

Dionne, M., and CL. Folt. 1991. An experimental analysis of macrophyte growth forms
as fish foraging habitat. Can. J. Fish. Aquat. Sci. 48: 123-131.

130

Downing, J .A. 1984. Sampling the benthos of standing waters, pp. 87-130 in Downing
J .A. and EH. Rigler eds. A manual on the methods for the assessment of
secondary productivity in fresh waters. 2Ind Edition. Blackwell Scientific
Publications, Boston.

Downing, J .A. 1986. A regression technique for the estimation of epiphytic invertebrate
populations. Freshwater Biology 16: 161-173.

Downing, J .A. and Cyr H. 1985. Quantitative estimation of epiphytic invertebrate
populations. Canadian Journal of Fisheries and Aquatic Sciences 42: 1570-1579.

Dvorak, J. and E.P.H. Best. 1982. Macro-invertebrate communities associated with the
macrophytes of Lake Vechten: structural and functional relationships.
Hydrobiologia 95:1 15-126.

Engel, S. 1990. Ecosystem responses to growth and control of submerged macrophytes: a
literature review. Madison, WI, Department of Natural Resources.

Engel, S. 1988. The role and interactions of submersed macrophytes in a shallow
Wisconsin lake. J. Freshwater Ecology 4(3): 329-341.

Englund, G., O. Samelle, S.D. Cooper. 1999. The importance of data-selection criteria:
Meta-analyses of stream predation experiments. Ecology 80(4): 1132-1141.

F assett, NC, 1957. A Manual of Aquatic Plants. The University of Wisconsin Press,
Madison, 405 pp.

Femandez-Duque, E. and C. Valeggia. 1993. Meta-analysis: A valuable tool in
conservation research. Conservation Biology 8(2): 555-561.

Galanti, G. 1995. A volumetric underwater mower (VUM) for hydrophytes assessment.
Comparison with alternative methods. Memorie dell’Istituto Italiano di
Idrobiologia Dott 53: 115-124.

Gaufin, A.R., E.K. Harris and H]. Walter. 1956. A statistical evaluation of stream
bottom sampling data obtained from three standard samplers. Ecology 37(4):
643-648.

Gerking, SD. 1957. A method of sampling the littoral macrofauna and its application.
Ecology 38(2): 219-226.

Gerrish, N. and J .M. Bristow. 1979. Macroinvertebrates associated with aquatic
macrophytes and artificial substrates. Journal of Great Lakes Research 5(1):
69-72.

131

Gilinsky, E. 1984. Role of fish predation and spatial heterogeneity in determining benthic
comrmmity structure. Ecology 65(2): 455-468.

Gregg, W.W. and F .L. Rose. 1985. Influences of aquatic macrophytes on invertebrate
community structure, guild structure, and microdistribution in streams.
Hydrobiologia 128: 45-56.

Gurevitch, J. and L.V. Hedges. 1999. Statistical issues in meta-analysis. Ecology 80(4):
1 142-1 149.

Haag, K.H. and GR. Buckingham. 1991. Effects of herbicides and microbial
insecticides on the insects of aquatic plants. J. Aquat. Plant Manage. 29: 55-57.

Hamelink, J .L., D.R. Buckler, F.L. Mayer and D.U. Palawski. 1986. Toxicity of
fluridone to aquatic invertebrates and fish. Environmental Toxicology and
Chemistry 5: 87-94.

Hanson, J .M. 1990. Macroinvertebrate size-distributions of two contrasting fresh-water
macrophyte communities. Freshwater Biology 24(3): 481-491.

Hargeby, A. 1990. Macrophyte associated invertebrates and the effect of habitat
permanence. Oikos 57: 338-346.

Hedges, L.V., J Gurevitch, and RS. Curtis. 1999. The meta-analysis of response ratios in
experimental ecology. Ecology 80(4): 1150-1156.

Jackson, MJ. 1997. Sampling methods for studying macroinvertebrates in the littoral
vegetation of shallow lakes. The Broads Authority, Norwich, 173 pp.

James, M.R., M. Weatherhead and E. Graynoth 1998. Macroinvertebrate distribution in
the littoral zone of Lake Coleridge, South Island, New Zealand - effects of habitat
stability, wind exposure, and macrophytes. New Zealand Journal of Marine and
Freshwater Research 32(2): 287-305.

J effries, M. 1993. Invertebrate colonization of artificial pondweeds of differing fractal
dimension. Oikos 67: 142-148.

Johnson, R.K. 1998. Spatiotemporal variability of temperate lake macroinvertebrate
communities: detection of impact. Ecological Applications 8(1): 61-70.

Keast, A. 1984. The introduced aquatic macrophyte, Myriophyllum spicatum, as habitat
for fish and their invertebrate prey. Can. J. Zool. Vol. 62: 1289-1303.

Kershner, M.W. and D.M. Lodge, 1990. Effect of substrate architecture on aquatic
gastropod-substrate associations. J. N. am. Benthol. Soc. 9(4): 319-326.

132

Komijow, R. 1989. Macrofauna of elodeids of two lakes of different trophy:
Relationships between plants and structure of fauna colonizing them. Ekologia
Polska 37(1-2): 31-48.

Kornikova, J. 1971. Quantitative relations between submerged macrophytes and
populations of invertebrates in a carp pond. Hydrobiologia 12: 377-3 82.

Krecker, RH, 1939. A comparative study of the animal population of certain submerged
aquatic plants. Ecology 20(4): 553-562.

Krull, J. N. 1970. Aquatic plant-macroinvertebrate associations and waterfowl. J.
Wildlife Mgrnt. 34: 707-718.

Lillie, R.A. and J. Budd. 1992. Habitat architecture of Myriophyllum spicatum L. as an
index to habitat quality for fish and macroinvertebrates. J. Fresh. Ecol. 7:
1 13-125.

Lyons, J. 1989. Changes in the abundance of small littoral-zone fishes in Lake Mendota,
Wisconsin. Can. J. Zool. 67: 2910-2916.

Madsen, J .D. 1991. Resource allocation at the individual plant level. Aquat. Bot. 41:
67-86.

Madsen, J .D. 1999. Point intercept and line intercept methods for aquatic plant
management. APCRP Technical Notes Collection (TN APCRP-Ml-OZ). US.
Army Engineer Research and Development Center, Vicksburg, MS.
www.wes.army.mil/el/aqua.

Madsen, J .D., J .W. Sutherland, J .A. Bloomfield, L.W. Eichler, and CW. Boylen. The
decline of native vegetation under dense Eurasian watermilfoil canopies. J. Aquat.
Plant Manage. 29: 94-99.

Madsen, J .D. and DH. Smith. 1997. Vegetative spread of Eurasian watermilfoil
colonies. J. Aquat. Plant Manage. 35: 63-68.

Madsen, J .D., L.W. Eichler and CW. Boylen. 1988. Vegetative spread of Eurasian
watermilfoil in Lake George, New York. J. Aquat. Plant Manage. 26:47-50.

Menzel, D. W., and N. Corwin. 1965. The measurement of total phosphorus in seawater
based on the liberation of organically bound fractions by persulfate oxidation.
Limnology and Oceanography 10:280-282.

Merritt, R.W. and K.W. Cummins, Eds. 1996. Aquatic Insects of North America.
Kendall/Hunt Publishing Co., Dubuque, Iowa. 862 p.

133

Merritt, R.W., J .R.Wallace, M.J. Higgins, M.K. Alexander, M.B. Berg, W.T. Morgan,
K.W. Cummins, and B. Vandeneeden. 1996. Procedures for the functional
analysis of invertebrate communities of the Kisimmee River-floodplain
ecosystem. Florida Scientist 59(4): 216-274.

Meyer, E. 1989. The relationship between body length parameters and dry mass in
running water invertebrates. Arch. Hydrobiol. 117(2): 191-203.

Mittelbach, G.G. 1981b. Foraging efficiency and body size: a study of optimal diet and
habitat use by bluegills. Ecology 62(5): 1370-1386.

Mittelbach, G.G. 1981a. Patterns of invertebrate size and abundance in aquatic habitats.
Canadian Journal of Fisheries and Aquatic Sciences 38: 896-904.

Mittelbach, G.G. 1988. Competition among refuging sunfishes and effects of fish density
on littoral zone invertebrates. Ecology 69(3): 614-623.

Mrachek, R.J., 1966. Macroscopic invertebrates on the higher aquatic plants at Clear
Lake, Iowa. Iowa Academy of Science 73: 168-177.

Muir, D.C.G., N.P. Grift, B.E. Townsend, D.A. Metner, W.L. Lockhart. 1982.
Comparison of the uptake and bioconcentration of fluridone and terbutryn by
rainbow trout and Chironomus tetans in sediment and water systems. 1982.
Environmental Contamination and Toxicology. 11:595-602.

Muir, D.C.G., B.E. Townsend, and W.L. Lockhart. 1983. Bioavailability of six organic
chemicals to Chironomus tetans larvae in sediment and water. Environmental
Toxicology and Chemistry. 2: 269-281.

Murphy, J ., and L. P. Riley. 1962. A modified single solution method for the
determination of phosphate in natural waters. Analytica Chimica Acta 27:31-36

Netherland, MD. and K.D. Getsinger. 1993. Control of Eurasian watermilfoil using
triclopyr. DowElanco Down to Earth. Vol. 48(1): 1-5.

Newman, R. M. 1991. Herbivory and detritivory on freshwater macrophytes by
invertebrates: a review. J. Am. Benthol. Soc. 10(2): 89-114.

Newroth, P. R. 1985. A review of Eurasian water milfoil impacts and management in
British Columbia. In: L.W.J. Anderson (ed), Proceedings of the First International
Symposium on the waterrrrilfoil (Myriophyllum spicatum) and related
Haloragaceae species. Aquat. Plant Manage. Soc., Washington, DC. pp. 139-153.

Nichols, D.S. and DR. Keeney. 1973. Nitrogen and phosphorus release from decaying
water milfoil. Hydrobiologia 42(4): 509-525.

134

Nusch, EA. 1980. Comparison of different methods for chlorophyll and phaeopigment
determination. Arch Hydrobiol. Beih. Ergeb. 14:14-36.

Olson, E. J ., E.S. Engstrom, M.R. Doeringsfeld, R. Bellig 1995. Abundance and
distribution of macroinvertebrates in relation to macrophyte communities in a
prairie marsh, Swan Lake, Minnesota. J. Freshwater Ecology 10(4): 325-335.

Olson, M.H., S.R. Carpenter, P.Cunningham, S. Gafrry, B.R. Herwig, N.P. Nibbelink, T.
Pellet, C. Storlie, A.S. Trebitz, and K.A. Wilson. 1998. Managing macrophytes to
improve fish growth: a multi-lake experiment. Fisheries 23(2): 6-12.

Painter, D.S. and K.J. McCabe. 1988. Investigation into the disappearance of Eurasian
watermilfoil from the Kawartha lakes. J. Aquatic Plant Management 26: 3-12.

Pardue, W. J. and D. H. Webb. 1985. A comparison of aquatic macroinvertebrates
occurring in association with Eurasian watermilfoil (Myriophyllum spicatum L.)
with those found in the open littoral zone. J. Freshwater Ecology 3(1): 69-79.

Parsons, J .K. and R.A. Matthews. 1995. Analysis of the associations between
macroinvertebrates and macrophytes in a freshwater pond. Northwest Science
69(4): 265-275.

Persson, L. and LB. Crowder. 1998. Fish-habitat interaction mediated via ontogenetic
niche shifts. The structuring role of submerged macrophytes in lakes. J eppeson,
E., M. Sondergard, and K. Christoferson Eds. New York, Springer-Verlag: 197-
213 (423 total).

Peterman, R.M., 1990a. The importance of reporting statistical power: the forest decline
and acidic deposition example. Ecology 71(5): 2024-2027.

Peterman, R.M., 1990b. Statistical power analysis can improve fisheries research and
management. Can. J. Fish. Aquat. Sci. 47: 2-15.

Pip, E. and J .M. Stewart. 1976. The dynamics of two aquatic plant-snail associations.
Can. J. Zool. 54: 1192-1205.

Pothoven, SA. 1996. An evaluation of the indirect effects of fluridone on the fish
communities in two Minnesota lakes. University of Minnesota MS. Thesis.
141 pp.

Pothoven, SA. and D.L. Vondracek. 1999. Effects of vegetation removal on bluegill

and largemouth bass in two Minnesota lakes. North American Journal of
Fisheries Management 19: 748-757.

135

Resh, V. H. I. and D. M. Rosenberg. 1979. Sampling variability and life history features:
basic considerations in the design of aquatic insect studies. J. Fish. Res. Board
Can. 36: 290-311.

Rogers, L.E., R.L. Buschbom and CR. Watson. 1977. Length-weight relationships of
shrub-steppe invertebrates. Ann. Ent. Soc. Am. 70(1): 51-53.

Rosine, W.N. 1955. The distribution of invertebrates on submerged aquatic plant surfaces
in Muskee Lake, Colorado. Ecology 36(2): 308-312.

Rosenberg, M.S., D.C. Adams, and J. Gurevitch. 1997. MetaWin: Statistical software for
meta-analysis with resarnpling tests. Sinauer Associates, Sunderling, MA 65 pp.

Rotenberry, J .T. and J .A. Wiens. 1985. Statistical power analysis and community-wide
patterns. Am. Nat. 125: 164-168.

Savino, J.F., E.A. Marschall and R.A. Stein. 1992. Bluegill growth as modified by plant
density: An exploration of underlying mechanisms. Oecologia 89: 153-160.

Schramm, H.L. Jr. and K.J. Jirka. 1989. Epiphytic macroinvertebrates as a food resource
for bluegills in Florida lakes. Transactions of the American Fisheries Society 118:
416-426.

Sheldon, SP. 1986. The effects of short-term disturbance of a freshwater macrophyte
community. J. Freshwater Ecology 3(3): 309-317.

Sher-Kaul, S., B. Oertli, E. Castella, J. Lachavanne. 1995. Relationship between biomass
and surface area of six submerged aquatic plant species. Aquatic Botany 51:
147-154.

Sloey, D., T. Schenck and R. Narf. 1997. Distribution of aquatic invertebrates within a

dense bed of Eurasian milfoil (Myriophyllum spicatum L.). J. Freshwater Ecology
12(2): 303-313.

Smith, CG. and J .W. Barko. 1990. Ecology of Eurasian Waternrilfoil. J. Aquat. Plant
Manage. 28: 55-64.

Smock, L.A., 1980. Relationships between body size and biomass of aquatic insects.
Freshwat. Biol. 10: 375-383.

Soszka, G.J., 1975. The invertebrates on submerged macrophytes in three Masurian
lakes. Ekol. P01. 23: 371-391.

Trebitz, A.S., S.A. Nichols, S.R. Carpenter, and R.C. Lathrop. 1993. Patterns of

vegetation change in Lake Wingra following a Myriophyllum spicatum decline.
Aquatic Botany 46: 325-340.

136

Trebitz, AS. and N. Nibbelink. 1996. Effect of pattern of vegetation removal on grth
of bluegill: a simple model. Canadian Journal of Fisheries and Aquatic Sciences
53: 1844-1851.

Vannote, R.L., G.W. Minshall, K.W. Cununins, J .R. Sedell, C.E. Cushing. 1980. The
river continuum concept. Can. J. Fish. Aquat. Sci. 37: 130-137.

Waters, T.F. 1969. Subsampler for dividing large samples of stream invertebrate drifi.
Limnology and Oceanography 14: 813-815.

Watkins, C. E. I., J .V. Sireman and W.T. Haller. 1983. The influence of aquatic
vegetation upon zooplankton and benthic macroinvertebrates in Orange Lake,
Florida. Aq. Plt. Mgrnt. 21: 78-83.

Welch, RS. 1948. Lirnnological Methods. McGraw-Hill, New York, 381 pp.

Werner, EB. and DJ. Hall. 1988. Ontogenetic habitat shifts in bluegill: the foraging
rate-predation risk trade-off. Ecology. 69(5): 1352-1366.

137

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

llIllllwrtawllrlrgwmwHI