NI

VIWIIIHWIN“WWI“LIIHLHIWNINN”

 

   

 

2007 LIBRARY I
Michigan State
University

 

 

 

This is to certify that the
thesis entitled

MACROINVERTEBRATE RESPONSE TO ZEBRA MUSSEL
(DREISSENA POLYMORPHA) COLONIZATION OF STREAM
SUBSTRATES

presented by

ALYSON ALISSA OLESEN

has been accepted towards fulﬁllment
of the requirements for the

degree In FISHERIES AND WILDLIFE

 

 

//%~

' Major Professor” 3 Signature

ngﬂpr- O '7-

Date

MSU is an aﬁinnative-action, equal—opportunity employer

_.—_—.-t-.-.--u----u-.--v- —._.—...._.—.-.-n-n-u-.-.-.-.-.-.-.-.-o--.-.-.-.-.-.-

. -.-.---—o-----o-.-._t— .. .-

 

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 DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/DateDue.indd-pt1

MACROINVERTEBRATE RESPONSE TO ZEBRA MUSSEL (DREISSENA
POLYMORPHA) COLONIZATION OF STREAM SUBSTRATES

By

Alyson Alissa Olesen

A THESIS
Submitted to
Michigan State University
In partial fulﬁllment of the requirements
For the degree of
MASTERS OF SCIENCE
Department of Fisheries and Wildlife

2007

ABSTRACT

MACROINVERTEBRATE RESPONSE TO ZEBRA MUSSEL (DREISSENA
POLYMORPHA) COLONIZATION OF STREAM SUBSTRATES

By
Alyson Alissa Olesen

Zebra mussels (Dreissena polymorpha) quickly spread throughout the Great
Lakes soon after being introduced to the region. They are now abundant in
many inland lakes, lake-outlet streams, and large rivers. The effects of
dreissenid colonization on native aquatic communities have been extensively
studied in lake habitats. However, unintentional spreading of dreissenids will
undoubtedly continue in other aquatic ecosystems, and it is important to
understand how lotic taxa may respond. I conducted an in-stream study of native
stream macroinvertebrate responses to simulated zebra mussel colonization
using shells adhered to ceramic tiles and a laboratory experiment to determine if
a prevalent taxon in the stream samples, hydropsychid larvae (Trichoptera),
exhibited substrate preference in the presence of live vs. dead zebra mussels.
My results suggest that only a few macroinvertebrate taxa exhibited positive
response to simulated zebra mussel colonization in streams. However, there was
a strong positive response by hydropsychid larvae to live zebra mussels in the
laboratory experiments. This suggests that benthic communities of small
streams are likely to become altered in response to spreading of dreissenids into
these ecosystems. Given that such macroinvertebrate responses have led to
signiﬁcant food web changes in lentic ecosystems, there is a great need for
management plans to both prevent and respond to future colonization in lotic

ecosystems.

ACKNOWLEDGEMENTS
Thanks to Oaklawn Hospital, Binder Park Zoo, and the Kellogg Experimental
Forest for allowing me to conduct my ﬁeld research on their properties. In
addition, I thank Reuben Goforth for his help with developing and executing the
project and Janice Sloan, Holly Campbell, Jen Pellegrini, and Fred Simmons who
helped collect or glue zebra mussels. I also thank Eric Benbow and Rich Merritt

for their guidance as my graduate committee members.

iii

TABLE OF CONTENTS

LIST OF TABLES ................................................................................... v
LIST OF FIGURES ................................................................................ vi
INTRODUCTION ................................................................................... 1
METHODS ............................................................................................ 7
RESULTS ........................................................................................... 14
DISCUSSION ...................................................................................... 28
SUMMARY ......................................................................................... 37
APPENDICES ...................................................................................... 39
REFERENCES .................................................................................... 43

iv

LIST OF TABLES

Table 1 .............................................................................................. 16
Table 2 .............................................................................................. 17
Table 3 .............................................................................................. 17
Table 4 .............................................................................................. 19
Table 5 .............................................................................................. 21
Table 6 .............................................................................................. 21
Appendix A .................................................................................... 39,40

INTRODUCTION

Zebra mussels (Dreissena polymorpha), native to the Ponto-Caspian
region, ﬁrst appeared in North America in Lake St. Clair (USA, Canada) in the
mid 19808 (Hebert et al. 1989). They quickly spread throughout the rest of the
Great Lakes, causing economic damage and ecological changes throughout the
basin (Stewart et al. 1998a, Perry et al. 2000). These changes result from the
zebra mussels ability to attach to almost anything sessile; they have essentially
become the substrate in many places in the Great Lakes, requiring expensive
removal for many industrial and recreational applications. They have also
displaced a pelagic driven food web to a benthic environment (Kuhns and Berg
1999, Lowe and Pillsbury 1995, Stewart and Haynes 1994). Mills et al. (2003)
referred to this shift to a benthic based food web as benthiﬁcation and suggests
that it has the potential to make benthic habitats more susceptible to the
establishment of other invasive species and to change the structure of native
benthic macroinvertebrate communities (Botts and Patterson 1996, Haynes et al.
1999, Haynes et al. 2005).

One of the most commonly observed ecological consequences of zebra
mussel colonization is an increase in benthic species richness and biomass in
response to increasing benthic primary production. Several factors contribute to
this increased biomass, including the interstitial habitats and refugia provided by
their shells, pseudofecal deposition (i.e., particles ﬁltered but not digested), and
water currents generated by their siphons which some macroinvertebrates take

advantage of for feeding (Ratti and Barton 2003, Ricciardi et al. 1997, Nalepa et

al. 2003, Mayer et al. 2001, Haynes et al. 2005). Shell complexity increases
habitat complexity that provides refugia from predators and wave action, and it
also traps phytoplankton and nutrients (Nitrogen and Phosphorus) that would
otherwise be swept away (Haynes et al. 1999, Stewart and Haynes 1994, Botts
and Patterson 1996). Zebra mussel feces and pseudofeces provide additional
food for benthic organisms (Vanderploeg et al. 2002, Greenwood et al. 2001).
Zebra mussels also increase water clarity by siphoning nutrients and
phytoplankton from the water column (Mayer et al. 2001, Lowe and Pillsbury
1995, Stewart and Haynes 1994). Greater water clarity allows deeper light
penetration, promoting benthic algal growth and concomitant increases in grazer
abundance. However, despite these positive inﬂuences, zebra mussels also
have negative effects on the through competitive displacement (Ricciardi et al.
1997, Stewart et al. 1998a). These responses by the benthos have been
reported based largely on observations in the Great Lakes, although the changes
are not restricted to these ecosystems.

Following colonization of the Great Lakes, zebra mussels have become
prevalent in many inland lakes, especially in the Great Lakes region, where
boaters readily move from lake-to-lake carrying veligers (larval zebra mussels)
on boat hulls, in bilges, and in live wells. As in the Great Lakes, these mussels
attach to any available substrates in inland lakes, including aquatic macrophytes,
and can dominate the benthic community (Horvath et al. 1999). Zebra mussels
have also increased water clarity in small lakes, which has increased macrophyte

growth and benthic invertebrate densities (Mayer et al. 2001). As zebra mussels

have successfully colonized isolated inland lakes, their colonization of lotic
systems is not unexpected.

Zebra mussels have entered large rivers in the United States, such as the
Ohio, Hudson, Huron, St. Lawrence, and Mississippi Rivers, as well as smaller
lake-outlet streams. The river and outlet stream zebra mussel populations are
often established ﬁrst by adults that are carried in by stream ﬂow or attached to
ﬂoating debris (Horvath et al. 1996, Horvath et al. 1999, Perry et al. 2000). The
mussels have not reached densities in rivers comparable to those observed in
lakes, perhaps because the moving river water creates high mortality for veligers
or decreases the ability of veligers to settle out and colonize substrates (Horvath
and Lamberti 1997). However, colonization of lotic habitats has occurred
nonetheless, suggesting the potential for continued spread into both large river
and smaller stream ecosystems.

Zebra mussels are often found just downstream from lake outlets because
of an increasing probability for veliger mortality as water ﬂow and turbulence
increase. However, Horvath and Lamberti (1997) suggested that macrophytes
are an important method for zebra mussel dispersal in rivers and streams. Adult
zebra mussels attached to macrophytes that are dislodged from the benthos or
cut up by boat trafﬁc can be carried into outflowing streams. These adults may
have a greater chance of colonizing the stream compared to veligers when the
macrophyte they are attached to becomes snagged in downstream areas. Adult
zebra mussels can then detach their byssal threads from the macrophytes and

move to more permanent stream substrates (Horvath and Lamberti 1997). This

suggests that there is a good likelihood for zebra mussels to become established
in smaller streams as they continue to spread.

There is a good potential for zebra mussels in lotic systems to induce
different responses by benthic invertebrates compared to benthic invertebrates of
lakes. First, the taxonomic composition of lentic communities is different from
lotic communities. In addition, zebra mussels enhance transfer of nutrients from
phytoplankton nutrients to benthic primary production (Greenwood et al. 2001).
Finally, nutrient rich feces or pseudofeces may be swept away by river currents
(Horvath et al. 1999). Therefore, habitat complexity, rather than the changes in
trophic structure, may be the primary driver for changes in macroinvertebrate
abundance when the zebra mussels are present in lotic ecosystems (Bolts and
Patterson 1996, Horvath et al. 1999, Stewart et al. 1998b).

While some evidence of response by benthic organisms in larger rivers
exists (e.g. Greenwood et. al. 2001, Strayer et. al. 1998, Haynes et. al. 1999, the
extent to which zebra mussels will colonize and inﬂuence smaller stream
ecosystems is poorly understood. There have been a few studies of
macroinvertebrate response in streams that already have zebra mussels (e.g.,
Greenwood et. al. 2001), but the potential effects of zebra mussels have not
been determined in headwater streams where they are not yet found.
Macroinvertebrate communities are already benthic driven in these systems, so
the effects zebra mussels will have on the community are unknown. However, it

is likely that benthic communities in small streams will respond largely to

increased habitat complexity rather than to displaced nutrients to the benthos
(Botts and Patterson 1996).

Among the benthic taxa in southern Michigan headwater streams,
collector-ﬁlterers are a group that may be expected to exhibit a response to the
structural complexity of zebra mussels. Organisms in the collector-ﬁlterer group
are suspension feeders and ﬁlter ﬁne particulate organic matter (FPOM) from the
water column (Merritt and Cummins 1996). Hydropsychid caddisﬂy larvae are
found in high abundances on rocks in ﬂowing streams (Mackay and Waters
1986) and belong to the collector-ﬁlterer functional feeding group (Wiggins 1996).
They build ﬁxed silken nets to capture drifting organic particles caught in the
stream current (Fairchild and Holomuzki 2002). Hydropsychid larvae were of
particular interest in this study because they were thought to beneﬁt from zebra
mussel colonization for both the habitat complexity onto which they can build
their nets and the siphoning action that could have increased their access to food
particles.

It is important to consider the ways in which stream macroinvertebrate
communities might respond to zebra mussel colonization for proactive
management. By studying the effects and responses prior to widespread
colonization, we may know what to expect when zebra mussels do colonize small
stream ecosystems, thus enabling the proactive development of management
plans. To address the potential management needs, my objective was to
observe aquatic insect larvae (i.e., changes in densities and taxa richness) to the

simulated presence of zebra mussels in headwater streams that have not yet

been colonized by dreissenids. I therefore used a combination of in situ and
laboratory experiments to quantify responses of stream benthic invertebrates to
the presence of zebra mussels to test several hypotheses related to benthic
macroinvertebrate responses in headwater streams.
Hypotheses
In situ Experiment
H1) Benthic macroinvertebrate densities and taxa richness will be greater on
zebra mussel shell tiles compared to natural substrate samples in southern
Michigan streams.
H2) Collector-ﬁlterer densities will be greater on zebra mussel shell tiles
compared to natural substrate samples in southern Michigan streams.
Laboratory Experiment
H3) Hydropsychid caddisﬂy larvae density will be:
a) greater on zebra mussel shell tiles vs. natural substrate tiles
b) greater on live zebra mussel tiles vs. natural substrate tiles

c) greater on live zebra mussel tiles vs. zebra mussel shell tiles

METHODS
In situ Experiment

Three headwater streams (Bear Creek, Calhoun 00., MI; Harper Creek,
Calhoun Co., MI; and Augusta Creek, Kalamazoo Co., Ml) that have not been
colonized by zebra mussels were used for in situ experiments and for collection
of hydropsychid larvae for lab experiments. All of these streams are tributaries of
the Kalamazoo River in southwestern Michigan and have similar riparian
characteristics (i.e., moderate to high riparian canopy cover, etc.). The substrate
in each stream was sandy with gravel and some cobble.

Zebra mussels collected in Lake Lansing (lngham Co., Ml) were emptied
and their shells were bleached using a 10% Chlorox® solution (The Chlorox Co.,
Oakland, CA). The valves were glued closed (Henkel LocTite® ControlTM Gel
Super Glue, Loctite Henkel Consumer Adhesive, lnc., Avon, OH) so that aquatic
insect larvae could not use the insides of the shells as habitat. A single layer of
zebra mussel shells was glued to terra cotta tiles (232 cm’) using PL®
polyurethane adhesive (Henkel, Mentor, OH). Eighty-one zebra mussel shells
were glued to each tile to achieve a density of 3521.7 individuals/m2. The density
of zebra mussel shells on each tile was based on previous studies on zebra
mussel densities in North America (e.g., Horvath et. al. 1996 and Ricciardi et. al.
1995), this experimental density was expected to elicit response by benthic

macroinvertebrates. The orientation of the zebra mussels on the tiles was

random. On 13 July 2006, nine tiles were placed in the run habitat of each of the
three streams to simulate the habitat complexity created by zebra mussel
colonization. Three tiles that had been covered with naturally occurring substrate
adhered with Gorilla Glue® (The Gorilla Glue Co., Cincinnati, OH) were also
placed in each stream at this time.

All tiles were retrieved on 29 September 2006, after an approximate
incubation period of 10 weeks. The incubation period was based on sample
considerations and time constraints for sampling. During tile retrieval, a D-frame
benthic sampling net (0.5 mm mesh size) was placed immediately downstream of
each tile to catch any larvae that became dislodged. All macroinvertebrates were
removed from the upper surface of the tiles and placed in 70% ETOH for later
laboratory processing. Representative community composition for naturally
occurring substrates in the study streams was also determined at the time the
tiles were retrieved by sampling randomly selected areas of natural substrate
using a 232 cm2 custom sample template. The steel template had an open top
and two solid sides. The template was placed on the stream bottom so that the
sides were parallel to the current and ﬁeld crew members used their hands to
disturb the substrate within the template area. Dislodged macroinvertebrates
were carried via stream ﬂow into a D-frame net placed directly downstream.
Sample Processing

Each sample was sorted and the macroinvertebrates were identiﬁed using
a dissecting microscope. The aquatic insects were identiﬁed to the genus level,

with the exception of chironomids which were identiﬁed to the family level, using

keys from Merritt and Cummins (1996). Other macroinvertebrates (e.g., leeches,
gastropods, limpets, oligochaetes) were identiﬁed to the lowest practical
taxonomic level based on time and resources. Mean total densities, functional
feeding group densities, and individual taxa densities were calculated as the
number of individuals present per square centimeter (no. individuals/cm?) and
then converted to the number of individuals present per square meter (no.
individuals/m2) for comparisons to other literature. Individual taxa identiﬁed from
the in situ experiment were assigned to appropriate functional feeding groups
(FFG) according to Merritt and Cummins (1996) and Smith (2001). FFGs reﬂect
the feeding mechanism of each taxon and provided a means by which rare taxa
not included in the individual taxa analyses because they were considered rare
could be included as part of the analyses. Placing taxa into FFGs also provided
information on the benthic macroinvertebrate community structure in the streams.
Taxa richness for each site was also determined as the numbers of individual
taxa identiﬁed in each sample.
Statistics

Statistical tests were conducted to determine whether benthic
macroinvertebrates responded to the introduction of zebra mussel shell tiles
(ZMST) at several levels. The treatments were split between 1) ZMST vs.
benthic samples to test for shell effects on macroinvertebrate colonization and 2)
benthic samples vs. natural substrate tiles to test for tile effects for the tile used
as a substratum for gluing the zebra mussel shells. A two-factor Analysis of

Variance (ANOVA) was initially used to determine if mean total densities of

macroinvertebrates and mean taxa richness measures were signiﬁcantly different
between treatments in the in situ experiment (a=0.05) (Version 15.0 2006, SPSS,
Chicago). Multivariate Analysis of Variance (MANOVA) tests were initially used
to test the hypotheses that the FFGs and individual benthic taxa densities
differed signiﬁcantly between the ZMST and benthic samples in the in situ
experiments (a=0.05) (Version 15.0 2006, SPSS, Chicago). The factors in this
test were stream and treatment, and the response variables were total densities
of benthic macroinvertebrate taxa, FFG densities, and individual taxa densities
observed on both the ZMST and in the benthic samples. All invertebrates
collected were included in the overall density and FFG analyses. However, only
taxa present in at least 20% of the samples taken from each stream were
included in the individual taxa analysis to exclude rare taxa that may have
skewed the results. Repeated efforts to normalize the data using transformations
were unsuccessful (e.g., log(x+1) and square root(x+1)). Thus, non-parametric
statistical methods were used in lieu of ANOVA and MANOVA to analyze the
ﬁeld study data.

The ﬁrst level of statistics performed tested for a stream-level effect.
Kruskal-Wallace (K-W) tests were used as a non-parametric equivalent of a two-
way ANOVA (o=0.05) (Version 15.0 2006, SPSS, Chicago). In cases where K-W
indicated a signiﬁcant difference among all streams, the Mann-Whitney (M-W)
test was used to determine signiﬁcant differences for each pair of streams in lieu
of a parametric post-hoc test (a=0.05). Lastly, densities found to be non

statistically different between a stream pair were tested for differences between

10

treatments using K-W tests (a=0.05). K—W tests were used to determine
statistical differences of mean total densities between treatments in each stream
(o=0.05).
Laboratory Experiment

Three recirculating-ﬂow laboratory aquaria (110 L) were used to observe
the response of hydropsychid caddisﬂy larvae to simulated colonization of zebra
mussels. Each aquarium was painted black on three sides using Quick Color
Spray Enamel (ROC Sales, Inc., Vernon Hills, IL). During the experiments, the
side left as a window was covered with construction paper so light could only
enter from the top of the aquarium. One end of the aquarium was raised using
plexiglass so that water would ﬂow in a downstream direction. The aquaria were
ﬁlled with water so that the shallow end was under approximately 5 cm of
deionized water. An aquarium power head (AquaClear® 3000, Rolf C. Hagen
Corp., Mansﬁeld, MA ) with a ﬂow rate of 757 L/h was placed in the sump to
pump water up through PVC pipes and down to the shallow end where it was ‘
returned via a spray bar (Figure 1). This simulated stream ﬂow over the shallow
end of the aquarium. The aquaria were used to house three experimental
treatments that provided resident hydropsychids with contrasting substrates for
colonization. Terra cotta tiles (232 cm2) were covered with the appropriate
substrate using Stick FastTM (Tech Marketing Inc., Peachtree City, GA) instant
adhesive in the same density as the in situ experiment (3521.7 individuals/m2).
They were placed on the bottom of the shallow side in each of the three

aquariums. The three experimental treatments used were: a) half zebra

11

mussels, half natural substrate, b) half zebra mussels, half empty zebra mussel
shells, and c) half empty zebra mussel shells, half natural substrate.

The hydropsychid caddisﬂy larvae used in the laboratory were collected in
Bear Creek (Calhoun Co., MI) and were allowed to acclimate to laboratory
temperatures for a minimum of 48 h before the experimental trials were
conducted. The number of hydropsychids used was based on ﬁeld observations
from natural substrate samples in the three study streams (i.e., 3 individuals/m2).
Twelve hydropsychids were placed in the center of the four tiles in each
aquarium and left to settle for 1 h before the water ﬂow from the pump was
turned on. Once the water ﬂow was initiated, the hydropsychids were left to
colonize the substrate for 48 h. At the end of the 48 h trial, the water ﬂow was
turned off and each tile was removed. The densities of hydropsychids on all tiles
were calculated to determine whether the hydropsychids colonized the
substrates differently. Any mortality was noted and a trial was rerun if more than
two hydropsychids were dead at the end of an experimental trial. The trials
were run twice, for a total of six replications per treatment pair. A one-way
ANOVA was used to determine whether hydropsychid densities varied among

the substrate treatments in the laboratory experiments (o=0.05) (SPSS 2006).

12

Live zebra

Substrate tile .
mussels tile

Live zebra Substrate tile
mu els tile

 

Figure 1. Diagram of lab experiment design and tile orientation. The power head
(a) pumped water through the PVC pipe (b) and out through a spray bar (c) to
allow the water out in a unidirectional ﬂow (indicated by arrows) across the tiles.

RESULTS
In situ Experiment

In total, 12 tiles (nine with zebra mussel shells and three with substrate)
were set and retrieved in each of the three experimental streams. In addition,
nine benthic samples were collected from natural substrates (hereafter referred
to as benthic samples) in each stream at the time that the tiles were retrieved.
Taxa richness was signiﬁcantly higher on ZMST compared to benthic samples in
Bear Creek (x 2=5.55, p=0.02). However, taxa richness in Augusta and Harper
Creeks was similar between treatments (x 2=0.72, p=0.40 and 1 2=3.32, p=0.07,
respectively) (Figure 2). Taxa richness was signiﬁcantly higher on substrate tiles
than in benthic samples in Harper Creek (x 2:429, p=0.04). However, taxa
richness in Augusta and Bear Creeks were not statistically different between
treatments (x’=2.21, p=0.14 and xz=0.93, p=0.33, respectively).

The overall mean macroinvertebrate densities on ZMST in each stream
were as follows: 1020.1 individuals/m2 (Augusta Creek), 6374.5 individuals/m2
(Harper Creek), and 450.2 individuals/m2 (Bear Creek) (Figure 3). The overall
mean macroinvertebrate densities in benthic samples in each of the streams
were as follows: 1230.8 individuals/m2 (Augusta Creek), 3486.6 individuals/m2
(Harper Creek), and 205.9 individuals/m2 (Bear Creek) (Figure 3). The most
common taxa across all streams included amphipods, elmids, oligochaetes, and
hydropsychids. These taxa were present in at least 20% of the samples taken

from each stream (Appendix A).

14

 

30

 

l Taxa Richness

 

20*

10“

Number of taxa

 

 

 

 

Bear Creek Augusta Creek Harper Creek

Figure 2. Taxa richness found in all samples in each experimental stream.

 

 

 

 

 

 

a

<

§ 10000 uzrvrsr

76 I Benthic sam les .
g 8000 ~ I Natural subs rate tile
E

g m 6000 d

V (D

a

'3 * 4000 -

C

O

'U

3 2000 -

3 _r_—_I—_,_I:E_,_
: 0 .

8

a

Bear Creek Augusta Creek Harper Creek
Figure 3. Mean total densities (+ 1 SE) of macroinvertebrates

(individuals/m2) in each of the three experimental streams by treatment.
ZMST vs. benthic samples

Mean total densities of macroinvertebrates were not statistically different
in Augusta and Harper Creeks between ZMST and in benthic samples (x’=0.24,
p=0.63 and x2=1.03, p=0.31, respectively). However, the mean total densities of
macroinvertebrates were signiﬁcantly higher on ZMST than in the benthic

samples in Bear Creek (x2=7.58, p=0.006) (Figure 3).

15

The collector-ﬁlterer group was the only functional feeding group with
densities that were statistically similar among all three streams (Table 1, Figure
4). Scraper densities were statistically similar between Augusta and Harper
Creeks (Table 1), and shredder densities were statistically similar between
Augusta and Bear Creeks (Table 1, Figure 4). No FFG groups were statistically
similar between Harper and Bear Creeks (Table 1).

K-W analyses were conducted to determine whether F FG densities were
different between ZMST and benthic samples based on the schedule of
statistically similar streams presented in Table 1. The results of this test
indicated that collector-ﬁlterer, scraper, and shredder densities were not
statistically different between the ZMST and benthic samples (26:036. p=0.55;
x’=0.17, p=0.68; 96:014. p=0.71, respectively) (Figure 4).

Table 1. Functional feeding groups with densities found to be statistically similar
between streams using Kruskal-Wallace (3;2 value, p-value) for among stream
comparisons and Mann-Whitney (Z-value, p-value) for painNise comparisons
(zebra mussel shell tiles vs. benthic samples). Streams are represented by (A)

Augusta Creek, (B) Bear Creek, and (H) Harper Creek. Collector-gatherer and
predator densities were not statistically similar between any of the stream pairs.

 

 

Test
Functional Feeding Group Streams value p-value
Collector-filterer A&H&B x’=4.76 0.09
Scraper A&I-I Z= -0.56 0.58
Shredder A38 Z= -0.02 0.99

 

The hydropsychid caddisﬂy group was the only taxon with densities that
were statistically similar among all three streams. All remaining taxa group
densities analyzed were based on comparisons between only two of the three

experimental streams according to the schedule provided in Table 2.

16

Table 2. Taxa with densities found to be statistically similar between streams
using Kruskal-Wallace ()6 value, p-value) for among stream comparisons and
Mann-Whitney (Z-value, p-value) for pairwise comparisons (zebra mussel shell
tiles vs. benthic samples). Streams are represented by (A) Augusta Creek, (B)
Bear Creek, and (H) Harper Creek.

 

 

 

 

 

 

 

Test p-

Class Subclass Order Family Streams value value
A&H, Z: -1.53, 0.13,
lnsecta Coleoptera Elmidae H&B Z= -1.51 0.13
Ephemeroptera Caenida A&H Z: -1.08 0.28
Heptageniidae A&H = -1.1 0.27
Trichoptera Hydropsychidae A&H&B X 2=3.66 0.16
Helicopsychidae H&B Z: -1.78 0.07
Leptoceridae A88 2: 0.00 1.00
Crustacea Amphipoda Gammaridae A&B Z= -0.88 0.38
Gastropoda A&B Z= -1.43 0.15
Clitellata Hirudinea A88 Z= 0.00 1.00

 

Table 3. Taxa found to be statistically similar between streams using Kruskal-
Wallace for among stream comparisons and Mann-Whitney for pairwise
comparisons and their statistical signiﬁcance between the treatments zebra
mussel shell tiles vs. benthic samples (X2 value, p-value). Streams are
represented by (A) Augusta Creek, (B) Bear Creek, and (H) Harper Creek.

 

 

 

 

 

 

Class Subclass Order Family Streams x2 p
A&H, 0.43, 0.51,
lnsecta Coleoptera Elmidae H88 3.73 0.05
Ephemeroptera Caenida A&H 1.68 0.19

Heptageniidae A&H 9.21 0.002
Trichoptera Hydropsychidae A&H&B 0.84 0.36
Helicopsychidae H&B 0.35 0.55
Leptoceridae A&B 0.00 1 .00
Crustacea Amphipoda Gammaridae A&B 0.78 0.38
Gastropoda A&B 0.002 0.97
Clitellata Hirudinea A&B 0.00 1.00

 

Densities of elmid beetle larvae were statistically similar between 1)
Augusta and Harper Creeks (Table 2) and 2) Harper and Bear Creeks (Table 2).
Elmid densities were not statistically different between ZMST and benthic
samples in Augusta and Harper Creek (Table 3) but were statically higher on

ZMST compared to benthic samples in Harper and Bear Creeks (Table 3, Figure

17

5). Caenid mayﬂy densities were similar between Augusta and Harper Creeks
(Table 2), but their densities were not statistically different between ZMST and
benthic samples (Table 3, Figure 5). Heptageniid mayﬂy densities were similar
between Augusta and Harper Creeks (Table 2), and their densities were
statistically higher on ZMST vs. benthic samples (Table 3, Figure 5).
Hydropsychid densities were similar between all streams (Table 2), but they were
not statistically different between ZMST and benthic samples (Table 3, Figure 5).
Helicopsychid densities were statistically similar between Harper Creek and Bear
Creek (Table 2), and were not statistically different between the treatments in
these streams (Table 3). Leptocerid densities were statistically similar between
Augusta and Bear Creeks (Table 2), and they not statistically different between
ZMST and benthic samples (Table 3). Amphipod, gastropod, and leech
densities were also statistically similar between Augusta Creek and Bear Creek
(Table 2), and the densities of all three of these taxa were not statistically
different between ZMST and benthic samples (Table 3).
Benthic samples vs. substrate tiles

Tests to determine whether the use of the tiles as a substrate for
simulating zebra mussel colonization inﬂuenced macroinvertebrate substrate
preference in the streams were based on nine benthic samples and three
substrate covered tiles from each stream. All the streams had statistically similar
mean total densities in benthic samples although they were statistically similar on
the substrate tiles in Augusta and Bear Creeks only (x2=1.92, p=0.17 and

xz=0.009, p=0.93, respectively) and the treatments were statistically higher on

18

substrate tiles compared to benthic samples in Harper Creek (x2=3.769, p=0.052)
(Figure 3).

The collector-ﬁlterer group was the only FFG with densities that were
statistically similar among all three streams (Table 4). Scraper densities were
statistically similar between Augusta and Harper Creeks (Table 4) and shredder
densities were similar between Augusta and Bear Creeks (Table 4, Figure 4). No
F FG groups were statistically similar between Harper and Bear Creeks (Table 4).

K-W analyses were conducted to determine whether FFG densities were
different between benthic samples and substrate tiles based on the schedule of
statistically similar streams presented in Table 4. Collector-ﬁlterer, scraper, and
shredder densities were not statistically different between the treatments
(36:006. p=0.81; x’=0.16, p=0.69; and x’=2.82, p=0.09, respectively).

Table 4. Functional Feeding groups with densities found to be statistically similar
between streams using Kruskal-Wallace (x2 value, p-value) for among stream
comparisons and Mann-Whitney (Z-value, p-value) for pairwise comparisons
(benthic samples vs. substrate tiles). Streams are represented by (A) Augusta

Creek, (B) Bear Creek, and (H) Harper Creek. Collector-gatherers and predators
were not statistically similar between any of the stream pairs.

Functional FeedingGroup Streams Test value p-value

 

Collector-ﬁlterer A&H&B X2: 4.63 0.10
Scraper A&H Z= -0.67 0.51
Shredder A&B Z: -0.06 0.95

 

Initial tests to determine whether resident taxa densities were different
among streams indicated that only densities of hydropsychid caddisﬂies and
heptageniid mayﬂies were statistically similar among all streams (Table 5). All
remaining taxa group densities analyzed were based on comparisons between

only two of the three experimental streams according to the schedule provided in

19

Table 5. However, no taxa had similar densities between Harper Creek and Bear
Creek (Table 5).

Densities of elmid beetle larvae were statistically similar between Augusta
and Harper Creeks (Table 5) and were also not signiﬁcantly different between
benthic samples and substrate tiles (Table 6, Figure 5). Caenid mayﬂy densities
were similar between Augusta and Harper Creeks (Table 5), but their densities
were not signiﬁcantly different between benthic samples and substrate tiles
(Table 6, Figure 5). Heptageniid mayﬂy densities were similar among all
streams, but they were not statistically different between the treatments (Table 6,
Figure 5). Diptera pupae densities were statistically similar between Augusta
Creek and Bear Creek (Table 5), but were not statistically different between
benthic samples and substrate tiles (Table 6, Figure 5). Hydropsychid densities
were similar among all streams, but they were not statistically different between
benthic samples and substrate tiles (Table 6, Figure 5). Leptocerid densities
were statistically similar between Augusta and Bear Creeks (Table 5), and they
were not statistically different between benthic samples and substrate tiles (Table
6). Amphipod, gastropod, leech, and oligochaete densities were also statistically
similar between Augusta Creek and Bear Creek (Table 5) and the densities of
these taxa were not statistically different between benthic samples and substrate

tiles (Table 6).

20

Table 5. Taxa found to be statistically similar between streams using Kruskal-
Wallace (x2 value, p-value) for among stream comparisons and Mann-Whitney
(Z-value, p-value) for painrvise comparisons (benthic samples vs. substrate tiles).
Streams are represented by (A) Augusta Creek, (B) Bear Creek, and (H) Harper
Creek.

 

 

 

 

 

 

 

 

 

Test p-
Class Subclass Order Family Streams value value

lnsecta Coleoptera Elmidae A&H Z: -1.65 0.1

Ephemeroptera Caenida A&H Z: -0.56 0.58

Heptageniidae A&H&B x2: 5.8 0.06

Diptera (pupae) A&B Z= -1.45 0.19

Trichoptera Hydropsychidae A&H&B X2: 0.96 0.62

Leptoceridae A&B Z= 0.00 1.00

Crustacea Amphipoda Gammaridae A&B Z: -0.79 0.43

Gastropoda A&B Z= -1.00 0.32

Clitellata Hirudinea A&B Z: 0.00 1.00

Oligochaeta A&B Z= -1.40 0.16

 

Table 6. Taxa found to be statistically similar between streams using Kruskal-
Wallace for among stream comparisons and Mann-Whitney for painrvise
comparisons and their statistical significance between the treatments benthic
samples vs. substrate tiles (x2 value, p-value). Streams are represented by (A)
Augusta Creek, (BLBear Creek, and (H) Harper Creek.

 

 

 

 

 

 

 

 

Class Subclass Order Family Streams x1 va'Iue
lnsecta Coleoptera Elmidae A&H 0.09 0.76

Ephemeroptera Caenida A&H 0.07 0.79

Heptageniidae A&H&B 0.001 0.98

Diptera (pupae) A&B 0.68 0.40

Trichoptera Hydropsychidae A&H&B 0.22 0.64

Leptoceridae A&B 0.00 1 .00

Crustacea Amphipoda Gammaridae A&B 0.32 0.57

Gastropoda A&B 0.33 0.56

Clitellata Hirudinea A&B 0.00 1.00

Oliﬂchaeta A&B 2.74 0.1

 

21

Functional Feeding Groups

 

 

 
   

DBear Creek a 1500 b
IAugusla Creek
1200‘
600‘.Harper Creek
m.-
400-

$

 

 

 

   

 

 

 

             

 

 

 

   

Mean density (individuals/m2) + SE

 

 

 

 

 

0‘ 0
8000 c 600 d
md
6000-
4m!
40004 300-
2m-
2000-
'°°" I I I
0 0 ZMST Benthic Substrate
samples tiles
150 e Treatment
120-
90-
60-
30-
0 ZMST Benthic Substrate
samples tiles
Treatment

Figure 4. Mean total densities (+ SE) of functional feeding groups for each
treatment in each stream, including: a) Shredders; b) Scrapers;

c) Collectors-gatherers; d) Collectors-ﬁlterers; e) Predators. Mean total den-
sity values used in graphs are means taken from the raw data, but statistical
tests were based on Kruskal-Wallace tests according to mean rank. ZMST=
zebra mussel shell tiles.

22

 

Individual Taxa

 

500:3 a CIAugusta Creel- m b

I Harper Creek
400 - I Bear Creek
I Total 150-

III)-1

1004

 

 

 

 

 

 

   

o

 

'ty (individuals/m2) + SE

Mean dens:

 

 

 

 

 

 

       

 

 

 

 

 

ZMST Benthic samples Substrate tiles 0 ZMST Benthic sanples Substraleliles
Treatment Treatment
Figure 5. Mean total densities (+ SE) for individual benthic taxa for each
treatment in each stream, including: a) elmidae; b) caenidae; c) heptageni-
idae; d) diptera pupae; e) chironomidae; f) hydropsychidae. Statistically
similar treatments are denoted by the same letters. Mean total density
values used in graphs are means taken from the raw data, but statistical
tests were based on Kruskal-Wallace tests according to rank. ZMST =
zebra mussel shell tiles.

   

 

23

Individual Taxa

 

9

ii

+SE
E

)

$5

 

 

 

o

 

Treatment

 

Mean density (individuals/m2
3 3 9

F

  

 

 

 

I _ .
ZMST BenthicSanples SubstaeTlle
Treatment

Figure 5 (cont’d). g) oligochaete; h) gammaridae; i) gastropoda (limpet).

24

 

Lab Experiment

Analyses of differences in hydropsychid caddisﬂy larvae densities among
live zebra mussels, zebra mussel shells, and simulated natural substrates were
based on data collected from three trials with each experimental comparison
replicated twice. The ﬁrst comparison tested to detect potential differences in
hydropsychid densities between ZMST similar to those used in the in situ
experiment and simulated natural substrate. Based on this analysis, there was
no signiﬁcant difference in hydropsychid densities between ZMST and the natural
substrate (F=1.11, p=0.31, df=1) (Figure 6).

There was no signiﬁcant difference in hydropsychid caddisﬂy densities on
the live zebra mussel tiles vs. the simulated natural substrate (F=0.66, p=0.43,
df=1) (Figure 6). However, there was a signiﬁcant interaction between treatment
and tile effects (i.e., upstream vs. downstream placement) (F=5.92, p=0.03,
df=1). There was no signiﬁcant difference in hydropsychid densities when an
ANOVA was conducted separately for both the upstream tiles (F=3.25, p=0.10,
df=1) and the downstream tiles (F=1.00, p=0.34, df=1). There was also an
interaction between the trial and tile effects (F=9.50, p=0.007, df=1). Based on
separate ANOVAs for each factor, there was no signiﬁcant difference in
hydropsychid densities between the treatments in the ﬁrst trial (F=0.39, p=0.55,
df=1). There was also no signiﬁcant difference in hydropsychid densities
between the treatments in the second trial (F=1.65, p=0.23, df=1).

The ﬁnal experimental trial tested for potential differences in hydropsychid

larvae densities between live zebra mussel tiles and ZMST. The results of this

25

experiment indicated that hydropsychid caddisﬂy densities were signiﬁcantly

higher on live zebra mussel tiles versus ZMST (F=5.51, p=0.03, df=1) (Figure 6).

26

 

0.03

0.02-

0.01-

 

 

 

 

 

ZMST Natural Substrate

 

.0
o
00

0.02-

.0

o

_s
I

 

 

   

 

Live zebra mussels Natural Substrate

Mean density (individuals/cm’) + SE

 

0.0"

0.024 *

0.01“

 

 

   

 

 

Live zebra mussels ZMST

Figure 6. Mean total densities (+ SE) of hydropsychid larvae for each
treatment in the laboratory experiment. *Densities were signiﬁcantly
different between treatments (p<0.05). ZMST=zebra mussel shell tiles.

27

DISCUSSION

The encroachment of dreissenids in North American freshwater
ecosystems is of great concern. Many studies have shown changes in benthic
macroinvertebrate community structure after the colonization of zebra mussels in
the Great Lakes. These studies have had contrasting results; some report that
densities of some benthic taxa increase in response to zebra mussel
colonization, while the densities of other benthic taxa decrease. Taken as a
whole, these results suggest that overall biomass and abundance of aquatic
invertebrates generally increase due to abiotic and/or biotic effects of zebra
mussels in the Great Lakes (Botts and Patterson 1996, Haynes et al. 1999,
Haynes et al. 2005, Kuhns and Berg 1999, Mayer et al. 2002, Reed et al. 2004,
Stewart and Haynes 1994, Stewart et al. 1998a, 1998b).

Studies of zebra mussel colonization in riverine and stream systems have
not been as common to date. Among the few existing studies, some have shown
that abiotic factors are more important than biotic factors in streams (Stewart et
al. 1998b, Horvath et al. 1999), while others show that biotic factors are just as
important (Greenwood et al. 2001). One component of the present study (the in
situ experiment) accounted for abiotic factors through the increased habitat
complexity provided by simulated zebra mussel colonization. The other
component of the study (the laboratory experiment) accounted for both biotic and
abiotic factors through the use of live zebra mussels. In both cases, the results

of this study provide some evidence to suggest that both biotic and abiotic factors

28

likely play roles in stream macroinvertebrate responses to zebra mussel
colonization.
In situ Experiment

Although each stream was chosen because of similar substrate and
riparian characteristics, it was apparent the benthic communities of the streams
were not as similar as initially expected after analyses were completed. For
example, overall benthic invertebrate densities were different between each
stream; with Harper Creek mean total densities much higher than the other two
streams. Chironomids were particularly abundant in the Harper Creek samples,
likely resulting from nutrient inputs from a zoo located upstream of the study site.
In contrast, Bear Creek mean total densities were much lower than the other two
streams, likely as a result of accumulated organic materials during tile incubation.
Agricultural ﬁelds located upstream of the site may have also provided greater
sediment loads to this stream compared to Harper and Augusta Creeks. Bear
Creek, however, had signiﬁcantly higher overall macroinvertebrate densities on
ZMST than in benthic samples. Oligochaetes, elmid larvae, and hydropsychid
larvae are three taxa that had much higher densities on ZMST than in the benthic
samples, which may explain the difference. These taxa may be responding to
the complexity created by the ZMST or the sediment that accumulated on the
tiles in Bear Creek. In contrast, Augusta Creek was located in a forest, but did
have agricultural and residential land upstream from the site; however, the site
used for the in situ experiment was much farther downstream of these land uses

compared to sites in the other two streams. Given that the landscape contexts of

29

the three experimental streams likely inﬂuenced the results of the in situ
experiment, the results presented herein should be interpreted with caution.
Regardless, the results are expected to provide a general indication of stream
benthic macroinvertebrate community responses to zebra mussel colonization in
a multi-land use landscape.

Only two FFGs were represented in Bear Creek - collector—ﬁlterers and
shredders. The presence of collector-ﬁlterers may be explained by the nutrient
inputs and sediment input from the agricultural lands upstream, and the presence
of shredders suggests that the stream has sufﬁcient allochthonous inputs to
provide support for this FFG. All FFGs were present in Augusta Creek; possibly
due to the absence of inputs from agricultural or residential land-uses
immediately upstream from the site increasing functional diversity in processing
organic matter. Similarly, all FFGs were present in Harper Creek, but with an
ovenrvhelming majority of collector-gatherers, especially chironomids. This is
likely due to inputs from a zoo upstream as previously discussed. The absence
of three of the FFGs (scrapers, collector-gatherers, and predators) in Bear Creek
once again suggests the experimental streams were dissimilar enough that the
landscape context of each stream was likely to be a signiﬁcant factor in
inﬂuencing the community structure of the streams, thus making experimental
comparisons using all the streams as replicates difﬁcult and often impossible.

The general lack of responses by stream benthic taxa to the introduction
of increased habitat complexity provided by the ZMST was surprising. Timing of

the experiment may have underestimated overall benthic macroinvertebrate

3O

response. However, this study was performed during the summer to avoid spring
and fall spate events that were expected to damage the experimental tiles. Taxa
such as amphipods, oligochaetes, and chironomids have been shown to have
greater responses to abiotic factors than biotic factors in lake systems (Botts and
Patterson 1996). Such abiotic factors (e.g., increased habitat complexity created
by zebra mussel shells) are thought to be the most important factors that
inﬂuence changes in benthic macroinvertebrate communities in the rocky
substrates of Lake Erie (Stewart et al. 1998b). However, increasing habitat
heterogeneity does not necessarily increase species richness (Wise and Molles
1979) or abundance, and the effect of zebra mussel colonization in streams has
been shown to be positive or negative depending on the characteristics of the
stream (Strayer et al. 1998, Haynes et al. 1999). The results of the in situ
experiment appear to concur with this in that the densities of only two taxa (elmid
larvae and heptageniid mayﬂies) were signiﬁcantly different between ZMST and
benthic sample treatments. In this case, the densities of both taxa were
signiﬁcantly higher on ZMST than in benthic samples. However, most other taxa,
including some groups that have shown responses to zebra mussel structure in
lentic habitats, did not exhibit a response to the increased habitat complexity
provided by the ZMST. This suggests that the increased structural complexity of
zebra mussel shells may not be a strong driver of benthic taxa in streams. Other
studies have increased habitat complexity by manipulating structures and have
demonstrated changes in stream benthic macroinvertebrate communities (e.g.,

Courtmanch 1984, Stewart et. al. 1998b, Obernborfer et. al. 1984). However, the

31

abiotic effect of zebra mussels may have been lessened in some cases in which
it appeared that the zebra mussels helped to trap inorganic substrate particles,
thus decreasing available habitat complexity. Studies in streams with lower
sediment loads may indicate different responses by benthic macroinvertebrates.
Use of live zebra mussels in the in situ experiments may have lessened the
accumulation of sediments on the tiles by their ﬁltering action reducing the
settling of inorganic and organic materials. However, the use of live zebra
mussels for this experiment was not an option for obvious reasons.

The increased densities of heptageniid mayﬂy nymphs on ZMST
compared to benthic samples be because the heptageniids responded to the
areas of exposed tile between the shells rather than the habitat complexity
created by the shells. Although no tile effect was found when comparing benthic
samples to substrate tiles, the tile covered with natural substrate did not have ﬂat
exposed areas comparable to those provided on the ZMST. Heptageniids are
clingers and are often found on and under loose cobble and boulders (Merritt and
Cummins 1996), which provide ﬂatter surfaces than those created by zebra
mussels. Heptageniids were also observed to be on the bottom sides of the tiles
when they were retrieved, and although these individuals were not included in
the samples, their presence suggests the preference of heptageniids for the ﬂat
surfaces provided by the tiles. While these exposed ﬂat surfaces may have
inﬂuenced heptageniid colonization, accumulated inorganic sediments on many
of the ZMST limited the amount of exposed tile on many of the ZMST similar to

the substrate tiles, suggesting that the heptageniids may have been relying

32

heavily on zebra mussel shell surfaces vs. exposed areas of tile. The use of
zebra mussel shells by heptageniids requires a more detailed behavioral study
that was beyond the scope of the present study.

Elmid beetle larvae also exhibited increased densities on ZMST compared
to benthic samples. Elmids, like heptageniids, are clingers and they are often
found in erosional and depositional areas (Merritt and Cummins 1996).
However, unlike the heptageniids, no elmids were observed to be clinging to the
bottom sides of the tiles during retrieval. Three possibilities exist to explain the
increased density of elmids on the ZMST: 1) elmids may be responding to the
inorganic sediments that accumulated on the ZMST during the incubation of the
tiles in the streams, 2) elmids may be clinging to the zebra mussel shell, and/or
3) elmids may be clinging to exposed tile. The use of zebra mussel shells by
elmids also requires a more detailed behavioral study.

The absence of biotic factors in the in situ experiment may have also
resulted in the general lack of response by stream benthic macroinvertebrates.
Biotic factors, such as higher food quality from the deposition of feces and
pseudofeces, have been found to be very important inﬂuencing factors on post-
zebra mussel colonization benthic macroinvertebrate communities (Ricciardi et
al. 1997, Stewart et al. 1998a, Greenwood et al. 2001, Mayer et al. 2002). While
some studies (Stewart et al. 1998b, Horvath et al. 1999) have found abiotic
factors to cause increases in macroinvertebrate density in zebra mussel
colonizations versus benthic samples, other studies have shown contrasting

results because feces and pseudofeces are likely to wash downstream in lotic

33

ecosystems. For example, a study by Horvath et al. (1999), which had both
zebra mussel shells and live zebra mussel treatments, found no signiﬁcant
difference in macroinvertebrate densities between the two treatments in their
stream. They suggested that this was because no accumulation of feces or
pseudofeces occurred as a result of stream ﬂow. In contrast, a study by
Greenwood et al. (2001) found that macroinvertebrate densities, especially
gastropods and amphipods, increased in the presence of zebra mussels because
of the higher food quality of their feces and pseudofeces in lotic systems.
Another study found that the only benthic macroinvertebrate to exhibit signiﬁcant
differences in density on areas of zebra mussel colonization were native unionoid
mollusks, which declined (Strayer et al. 1998). The in situ experiment focused
only on abiotic factors inﬂuencing macroinvertebrate density changes on
simulated zebra mussel colonization due to my inability to use live zebra mussels
as part of the in situ experiment. While the interstitial spaces did allow for
organic debris to fall out and collect on the tiles, this type of organic debris may
not have been of sufﬁcient nutritional value to generate the kinds of responses
that zebra mussel feces and pseudofeces elicit.

Densities of zebra mussels have been reported to be 400,000 mussels/m2
in Lake Erie (Maclssac et al. 1991), 60,000 mussels/m2 near the headwaters of
the Rhine River in Europe (Cleven and Frenzel 1993), and 100 mussels/m2 in the
St. Joseph River in southwestern Michigan (Horvath et al. 1996). It has been
suggested that zebra mussels must exceed a density of 1000 mussels/m2 to

signiﬁcantly affect unionid clams in the St. Lawrence River (Ricciardi et al. 1995).

34

In this study, a simulated density of 3521.7 individuals/m2 was used to test the
effects of zebra mussel colonization on benthic densities. While this density has
not been proven to be sufﬁcient to generate a response by stream
macroinvertebrates, it was nonetheless expected to have had an effect on the
macroinvertebrates in the streams. Additional studies should be performed to
consider the threshold density in which macroinvertebrates are affected by zebra
mussel colonization.
Lab Experiment

The laboratory experiment tested the potential effects of both biotic and
abiotic factors on hydropsychid larvae substrate preferences because both live
zebra mussel tiles and ZMST treatments were used. Hydropsychid larvae prefer
stable substrates, such as large stones or logs, in high ﬂow areas and tend to
avoid settling in less stable substrates such as ﬁne gravel (Fairchild and
Holomuzki 2002). Further, they prefer habitats that are structurally complex to
provide refuge from predators, materials useful in retreat construction, and easy
access to foraging (Fairchild and Holomuzki 2002). Hydropsychid caddisﬂy
larvae also prefer to inhabit areas of higher turbulence, such as notches in rocks
and boulders, because it increases the efﬁciency of prey being caught in their
nets (Osborne and Herricks 1987, Hart and Finelli 1999). Hydropsychid larvae
are also territorial and the increased habitat complexity provided by zebra mussel
colonization may allow greater numbers of hydropsychids to coexist. In general,
as velocity and/or food concentration increases, the distance in territorial

caddisﬂies decreases (Matczak and Mackay 1990). Thus, hydropsychids could

35

respond positively to both the increases in microturbulence and/or the potential
increase in food availability resulting from zebra mussel colonization, and the
simulated colonization by zebra mussels was expected to inﬂuence all of these
requirements in the laboratory experiment.

Changes in substrate complexity (i.e., interstitial spaces and notches)
change the microhabitat characteristics (Osborne and Herricks 1987). The
changes in substrate and microvelocity zebra mussels provide via interstitial
spaces may inﬂuence responses of some taxa to prefer to inhabit zebra mussel
colonies. Filter-feeders, such as hydropsychids, are commonly found in areas
with high velocity (Hart and Finelli 1999). In this study, the interstitial spaces
among the zebra mussels would increase the turbulence. The hydropsychid
larvae inhabited tiles with live zebra mussels rather than dead zebra mussels but
not necessarily in other treatments. The results, however, would most likely be
more prominent in natural zebra mussel colonies in streams because natural

colony densities, mussel positions, and clumping were not replicated on the tiles.

36

SUMMARY

Although there was limited response by macroinvertebrates in the in situ
experiment based on overall densities, functional feeding group assignments, or
individual taxa to the simulated colonization of zebra mussels, the laboratory
experimental results suggest that live zebra mussels may more strongly inﬂuence
macroinvertebrates in streams. In the presence of both abiotic and biotic factors
provided by live zebra mussels, a representative taxon (i.e., hydropsychid
caddisﬂy larvae) chose to inhabit substrate with live zebra mussels rather than
substrate with dead zebra mussels. Quagga mussels (Dreissena bugensis) are
a close relative of the zebra mussels. They are also native to the Ponto—Caspian
area and have been recently found spreading in the Great Lakes region. Similar
results are expected in response to quagga mussel invasion of streams.
Additional studies are needed to determine the responses of benthic
macroinvertebrates to the colonization of dreissenids in different habitats and in
streams with different characteristics. In addition, the speciﬁc behavioral
responses of certain taxa (i.e., heptageniids, elmids, and hydropsychids) may
provide greater insight into the potential inﬂuences of zebra mussel colonization
on stream benthos. Regardless, my study suggests at least some level of
response by headwater stream benthic macroinvertebrates, indicating a need for
the development of management plans to address future introductions into
headwater streams. Further, based on the differences in benthic community

structure among the three streams used in my study, it is likely that such

37

management approaches will need to be adaptive depending on stream

characteristics, from local riparian conditions to larger scale basin land use.

38

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 
 

0.0:... x 0.33.04.32.00
..2o0_.oo
0000.50 x 002530.83
0000.80 x 080002002 00200200..
0000.80 x 08200200.. 002.082.80.00.
0.2.... x 8080.09.00: 000.82.000.02.
..2o0__oo
0.2.... x x x 0003018053000 0028000902..
..2o0__oo
000.00 x x 050008.00: 00250820...
0.0500 x 08.000302m 002.800.8005 0.0.02.2...
..2o0._oo
'33. x 0.0.40 08.0.0 as. 20 s.
.0000. :0 x 0380.0 002.030 3%
0.0501 x 38...... 0%»
..2o0__oo
.080...» x 030.: 00230:
20020 x 0:25 0028003
0.0500 x 009000.33. 000382.85
..200..8
0.0500 x x x 002822.80
..200..8
2000.0 x x 0.50080 0025082000 0.0.0.0
000.00 x 04.230
000.00 x 00020.”. 0028080000
000.00 x x x 23.000000 000.8.m
0.0500 x x 030000.05 000.8.m 0.000330 0.000... 00o0o......<
..2o0__oo
I220 .320 Jﬁ. Ixoﬂwl Issue 3.50. .85 323:0 320 E... :0

880.35“. .03 .000... 50:0:<

 

800.50

0:20 0.80888...

 

.80020 8000 .2 00.0800 05 .o «com :05 0.08:. 8000.0

0.0. 002: 8000.00. 0.2. c. :20... 8088000 0:20 08.000. 082.082 080 80020 002800. 0x0... .< x.0:000<

39

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.208... x x 21:802.»: 82502
Jam. x 00.00300-
.0285 x 028.88 82:8,. 3
000020 x x x 03.08800 000008800 3 0000.080 00000.8(
080.5 x x x g
2000.0 x 008.08.... 08:0...0 02.08.<
000.8 x x magma.
.20.... x x 0_I>_u>|_m 802.02
..200:00
I008... x 0.00%....
000020 x 08080.20 0318020 3
lag x 0:: 8000.... O 0002 800
20020 x 02:000.. 00020800
20020 x 0000800208.“. 00020800
.2000.0 x 00.800808... 002:0..008000 0.08000
I403 x x 0%
..200__00
.000.00 x x 00.00820 002200200...
.000.00 x 082.0020 002800080...
.0.0500 x x 0.8000 0002000
..200:00
.0.0500 x 000.0000 000.00..00m
..200_.00
.0.0500 x 002.00m 0.0.00.08020m 0000... 00000.8(
..200:00
Idﬂloll gloom 4000' Ixoddl 2....0 8...... 3.23:0 .220 8.620 820 :5. i

80000.5“. .000 .008... 30:93

 
 

 

800.5

030.0 0.82.0.0...

 

.80050 2000 .0. 00.0800 05 .0 axoom :05 0.08:. 8000.0 0.0. 0005
8000.00. 0.00 :. 80x0... 80800.000 0:0.0 08.000. .0..0..0..:. 080 .80050 .00....802 0x0... .2800. < x.0:000<

4O

Appendix B. In situ experiment photos (a-f) and laboratory experiment photos
(g-h). a) Augusta Creek; b) Harper Creek; c) Bear Creek; d) zebra mussel shell
tile (ZMST); e) ZMST incutbating in stream; f) tile retrieval; 9) laboratory experi-
ment design; h) laboratory tile orientation.

 

41

Appendix 8 (continued). 9) laboratory experiment design; h) laboratory tile
orientation.

 

42

REFERENCES

Botts, PS, and BA. Patterson. 1996. Zebra mussel effects on benthic
invertebrates: physical or biotic? Journal of the North American
Benthological Society. 15(2): 179-184.

Cleven, E.J., and P. Frenzel. 1993. Population dynamics and production of
Dreissena polymorpha (Pallas) in River Seerhein, the outlet of Lake
Constance (Obersee). Archiv fur Hydrobiologie. 127: 395-407.

Courtmanch, D.L. 1984. A closing artiﬁcial substrate device for sampling benthic
macroinvertebrates in deep rivers. Freshwater Invertebrate Biology. 3(3):
143-146.

Fairchild, MP, and JR. Holomuzki. 2002. Spatial variability and assemblage
structure of stream hydropsychid caddisﬂies. Journal of the North
American Benthological Society. 21(4): 576-588.

Greenwood, K.S., J.H. Thorp, R.B. Summers, and D.L. Guelda. 2001. Effects of
an exotic bivalve mollusc on benthic invertebrates and food quality in the
Ohio River. Hydrobiologia. 462: 169-172.

Hart, DD. and CM. Finelli. 1999. Physical-biological coupling in streams: the
pervasive effects of ﬂow on benthic organisms. Annual Review of Ecology
and Systematics. 30: 363-395.

Haynes, J.M., T.W. Stewart, and GE. Cook. 1999. Benthic macroinvertebrate
communities in southwestern Lake Ontario following invasion of
Dreissena: continuing change. Journal of Great Lakes Research. 25(4):
828-838.

Haynes, J.M., N.A. Tisch, C.M. Mayer, and RS. Rhyne. 2005. Benthic
macroinvertebrate communities in southwestern Lake Ontario following
invasion of Dreissena and Echinogammarus: 1983 to 2000. Journal of the
North American Benthological Society. 24(1): 148-167.

Hebert, P.D.N., B.W. Muncaster, and G.L. Mackie. 1989. Ecological and genetic
studies on Dreissena polymorpha (Pallas): a new mollusk in the Great
Lakes (USA, Canada). Canadian Journal of Fisheries and Aquatic
Sciences. 46: 1587-1591.

Horvath, T.G., G.A. Lamberti, D.M. Lodge, and W.L. Perry. 1996. Zebra mussel
dispersal in lake-stream systems: source-sink dynamics? Journal of the
North American Benthological Society. 15(4): 564-575.

Horvath, T.G., G.A. Lamberti. 1997. Drifting macrophytes as a mechanism for

43

zebra mussel (Dreissena polymorpha) invasion of lake-outlet streams. The
American Midland Naturalist. 138(1): 29-36.

Horvath, T.G., K.M. Martin, and GA. Lamerti. 1999. Effect of zebra mussels,
Dreissena polymorpha, on macroinvertebrates in a lake-outlet stream. The
American Midland Naturalist. 142(2): 340-347.

Kuhns, L.A., and MB. Berg. 1999. Benthic invertebrate community responses to
round goby (Neogobius melanostomus) and zebra mussel (Dreissena
polymorpha) invasion in southern Lake Michigan. Journal of Great Lakes
Research. 25(4): 910-917.

Lowe, R.L., and R.W. Pillsbury. 1995. Shifts in benthic algal community structure
and function following the appearance of zebra mussels (Dreissena
polymorpha) in Saginaw Bay, Lake Huron. Journal of Great Lakes
Research. 21(4): 558-566.

Maclsaac, H.J., W.G. Sprules, and J.H. Leach. 1991. Ingestion of small-bodied
zooplankton by zebra mussels (Dreissena polymorpha): can cannibalism
on larvae inﬂuence population dynamics. Canadian Journal of Fisheries
and Aquatic Sciences. 48: 2051-2060.

Mackay, R.J. and T.F. Waters. 1986. Effects of small impoundments on
hydropsychid caddisﬂy production in Valley Creek, Minnesota. Ecology.
67(6): 1680-1686.

Matczak, T2. and R.J. Mackay. 1990. Territoriality in ﬁlter-feeding caddisﬂy
larvae: laboratory experiments. Journal of the North American
Benthological Society. 9: 26-34.

Mayer, C.M., L.G. Rudstam, E.L. Mills, S.G. Cardiff, and CA. Bloom. 2001.
Zebra mussels (Dreissena polymorpha), habitat alteration, and yellow
perch (Perca ﬂavescens) foraging: system-wide effects and behavioural
mechanisms. Canadian Journal of Fisheries and Aquatic Sciences. 58:
2459-2467.

Merritt, R.W., and K.W. Cummins, ed. 1996. An Introduction to the Aguatic
Insects of Ngrth America, 3'd ed. Iowa: Kendall/Hunt Publishing Co.

Mills, E.L., J.M. Casselman, R. Dermott, J.D. Fitzsimons, G. Gal, K.T. Holeck,
J.A. Hoyle, O.E. Johannsson, B.F. Lantry, J.C. Makarewicz, E.S. Millard,
l.F. Munawar, M. Munawar, R. O’Gorrnan, R.W. Owen, L.G. Rudstam, T.
Schaner, and T.J. Stewart. 2003. Lake Ontario: food web dynamics in a
changing ecosystem (1970-2000). Canadian Journal of Fisheries and
Aquatic Sciences. 60: 471-490.

44

Nalepa, T.F., D.L.Fanslow, M.B. Lansing, and GA. Lang. 2003. Trends in the
benthic macroinvertebrate community of Saginaw Bay, Lake Huron, 1987
to 1996: responses to phosphorus abatement and the zebra mussel,
Dreissena polymorpha. Journal of Great Lakes Research. 29(1): 14-33.

Overnborfer, R.Y., J.V. McArthur, J.R. Barnes. 1984. The effect of invertebrate
predators on leaf litter processing in an alpine stream. Ecology. 65(4):
1325-1331.

Osborne, LL. and E.E. Herricks. 1987. Microhabitat characteristics of
Hydropsyche (T richopter:Hydropsychidae) and the importance of body
size. Journal of the North American Benthological Society. 6(2): 115-124.

Perry, W.L., D.M. Lodge, and GA. Lamberti. 2000. Crayﬁsh (Orconectes
rusticus) impacts on zebra mussel (Dreissena polymorpha) recruitment,
other macroinvertebrates and algal biomass in a lake-outlet stream. The
American Midland Naturalist. 144: 308-316.

Ratti, C. and DR. Barton. 2003. Decline in the diversity of benthic invertebrates
in the wave-zone of eastern Lake Erie, 1974-2001. Journal of Great Lakes
Research. 29: 608-61 5.

Reed, T., S.J. Wilgus, A.K. Barnes, J.J. Schiefelbein, and AL. Fettes. 2004.
Refugia and local controls: benthic invertebrate dynamics in lower Green
Bay, Michigan following zebra mussel invasion. 30(3): 390-396.

Ricciardi, A., F.G. Whoriskey, and J.B. Rasmussen. 1995. Predicting the intensity
and impact of Dreissena infestation on native unionid bivalves from
Dreissena ﬁeld density. Canadian Journal of Fisheries and Aquatic
Sciences. 52: 1449-1461.

Ricciardi, A., F .G. Whoriskey, and J.B. Rasmussen. 1997. The role of the zebra
mussel (Dreissena polymorpha) in structuring macroinvertebrate
communities on hard substrata. Canadian Journal of Fisheries and
Aquatic Sciences. 54: 2596-2608.

Smith, 0.6. 2001. Pennak’s Freshwater Invertebrates of the gnited States:
porifera to crustacea, 4th ed. New York: John Wiley & Sons, Inc.

Statistical Package for the Social Sciences (SPSS). 2006. Version 15.0 for
Windows. SPSS Inc., Chicago, IL.

Stewart, T.W., J.M. Haynes. 1994. Benthic macroinvertebrate communities of
southwestern Lake Ontario following invasion of Dreissena. Journal of
Great Lakes Research. 20(2): 479-493.

45

Stewart, T.W., J.G. Miner, and R.L. Lowe. 1998a. Macroinvertebrate
communities on hard substrates in western Lake Erie: structuring effects
of Dreissena. Journal of Great Lakes Research. 24(4): 868-879.

Stewart, T.W., J.G. Miner, and R.L. Lowe. 1998b. Quantifying mechanisms for
zebra mussel effects on benthic macroinvertebrates: organic matter
production and shell-generated habitat. Journal of the North American
Benthological Society. 17(1): 81-94.

Strayer, D.L., L.C. Smith, and DC. Hunter. 1998. Effects of the zebra mussel
(Dreissena polymorpha) invasion on the macrobenthos of the freshwater
tidal Hudson River. Canadian Journal of Zoology. 76(3): 419-425.

Vanderploeg, H.A., T.F. Nalepa, D.J. Jude, E.L. Mills, K.T. Holeck, J.R. Liebig,
|.A. Grigorovich, and H. Ojaveer. 2002. Dispersal and emerging ecological
impacts of Ponto-Caspian species in the Laurentian Great Lakes.
Canadian Journal of Fisheries and Aquatic Sciences. 59: 1209-1228.

Wiggins, GB. 1996. Trichoptera Families. In An Introduction Lg the Aquatic
Insects of North America_, 3'd ed. Merritt, R.W. and K.W. Cummins, eds.
Iowa: Kendall/Hunt Publishing Co., 1996.

Wise, DH. and MC. Molles, Jr. 1979. Colonization of artiﬁcial substrates by

stream insects: inﬂuence of substrate size and diversity. Hydrobiologia.
65:69-74

46

 

Iill"HilllLlllleljjlllilfjllfI