Using Multiple Sentinel Species and Stable Isotopes to Understand Mercury Sources and Fate in Temperate Streams

USING MULTIPLE SENTINEL SPECIES AND STABLE ISOTOPES TO UNDERSTAND MERCURY SOURCES AND FATE IN TEMPERATE STREAMS

Timothy D. Jardine

B.Sc., Dalhousie University, 1996-2000 M.Sc., University of New Brunswick, 2001-2003

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

Doctor of Philosophy

In the Graduate Academic Unit of Biology

Supervisor(s): Dr. Karen A. Kidd (Biology)

Examining Board: Dr. Jeff Houlahan (Biology) Dr. James Kieffer (Biology) Dr. Charles Bourque (Forestry and Environmental Management)

External Examiner: Dr. Gilbert Cabana (Departément de Chimie-Biologie, Université du Québec à Trois Rivières)

This thesis is accepted by the Dean of Graduate Studies.

THE UNIVERSITY OF NEW BRUNSWICK

December, 2008

©Timothy D. Jardine, 2009

GENERAL ABSTRACT

The goal of this dissertation was to better understand mercury (Hg) dynamics in

running waters. It employed two techniques, analysis of stable isotopes of carbon and nitrogen and the use of sentinel species, to achieve this goal. I began by learning about the feeding ecology of a common predatory insect, the water strider (Aquarius remigis) that I hoped to use as an indicator of aquatic Hg contamination. I found that water striders exhibited a strong connection to the riparian zone via consumption of terrestrial insects that made up a majority of their diet. Water striders also had Hg concentrations that were weakly related to variables known to influence Hg in waterbodies, such as pH and dissolved organic carbon content. Instead water strider Hg concentrations were predicted by their body size, trophic level and proximity to a coal-fired power plant that

annually emits ~ 100 kg of Hg to the atmosphere. This suggests that striders are

terrestrial organisms that happen to spend much of their time on the surface of the water,

and they tell us little about processes occurring in the aquatic environment.

I then sampled algae, invertebrates and fishes from 60 stream and river sites in

New Brunswick, Canada from 2004-2007. I found that the small minnow, blacknose dace, accumulated far more Hg than the larger species brook trout that is often caught by

recreational anglers. Mercury concentrations in blacknose dace were predicted mainly by

the pH of the water; acidic streams had dace with highest Hg concentrations. Trout Hg

concentrations were determined more by their diet and their proximity to the coal-fired

power plant. Trout that fed higher on the food chain and lived in streams within 50 km of

the power plant had the highest Hg concentrations, but none of these concentrations were

above the human health consumption guideline of 0.5 ug g-1 wet weight. Mercury

increased four- to five-fold with each step in the food chain, similar to rates found in lakes and oceans, and differences in Hg concentrations in fishes were determined largely by differences at the base of the food chain.

All of these analyses suggest that patterns of Hg biomagnification in streams and rivers is similar to that observed in other ecosystems, and that Hg exposure will depend on the species that are present and the amount of Hg available at the base of the food chain.

iii

ACKNOWLEDGEMENTS

A thesis is never a product of a single person, and this work is no exception.

Given that the ideas for this dissertation began as early as 2003, and it is now 2008, it is no surprise that I have had plenty of help along the way.

I’ll start with my supervisor Karen. I consider her to be the ideal supervisor, no mean feat since she is so new at this game. At times she was thoughtful and caring (e.g. insisting on celebrating birthdays even if she was made aware months too late). At other times she was tough and decisive, insisting on a high-quality product no matter the consequences. I am forever grateful that she was willing to take me on in the first place and then allow me to leave before the end of three years to start a post-doc in Australia.

Thank you Karen.

Next up is my committee: Rick, Kelly and Neil. Where would I be without them?

Perhaps I would be climbing telephone poles to try and collect osprey feathers. Did I really think I could do that? Thanks for keeping me on track so far.

A lot of the work I did during this degree involved solo missions, driving alone to remote points in New Brunswick, not telling anyone where I had gone or when to expect me back. However there was also no shortage of teamwork required, particularly when it came time to do some electro-fishing. I am grateful to have had numerous individuals who gave their time to spend with me chasing fish and those elusive water striders. These include: Aaron Fraser, Matt Sabean, Sherisse McWilliam, Tim

Arciszewski, Kelly Lippert, Tim Barrett, Matt Sullivan, Rutger Engelbertink, Philip

Brett, Noel Swain, Sara Fraser, Len Giardi, Pete Emerson, Craig Poole, Rachel Keeler, and Chris Blanar.

I also had plenty of assistance with laboratory analyses. These analyses ranged

from water quality (John O’Keefe) to total (Paul Arp, Dragana Perkmann, Duc Banh, and

Mina Nasr) and methyl (David Lean, Emmanuel Yumvihoze, Ogo Nwobu, Brianna Wyn,

Leanne Baker, Ellen Belyea and Ellen Campbell) mercury analysis and most of all, stable isotopes. One of the advantages of working in an isotope lab is the all-access privileges that you acquire. Anne, Mireille and Christine had the patience to deal with my incomplete submissions, last minute requests and other incessant demands as I scrambled to get the data. I’ve learned a lot from each of you.

I was generously funded by several agencies during this project. NSERC came through with a postgraduate scholarship to me and Discovery Grant funding to Karen.

The New Brunswick government has been on board from the start, providing funds from

WTF and ETF. The O’Brien Humanitarian Trust gave me a large sum of money early on in the project. Lastly UNB graciously kicked in with Grand Lake Meadows Fund money, as well as a Doctoral Tuition award and an Eaton Fellowship for travel.

Finally I must thank my wife Laura. When I started at UNB so many years ago, I had no idea I would be lucky enough to have such a wonderful person with whom to share my life. It also doesn’t hurt that she was a source of skilled labour.

And Tinky helped me too.

Table of Contents

GENERAL ABSTRACT...... ii ACKNOWLEDGEMENTS...... iv Table of Contents...... vi List of Tables ...... viii List of Figures...... ix Chapter 1.0 Introduction to the Problem ...... 1 1.1 Background...... 1 1.2 References...... 9 Chapter 2.0 Applications, Considerations, and Sources of Uncertainty When Using Stable Isotope Analysis in Ecotoxicology ...... 16 Abstract...... 16 2.1 Introduction...... 17 2.1.1 Qualitative Studies...... 19 2.1.2 Biomagnification Studies...... 20 2.1.3 Quantitative Assessment of Carbon Sources ...... 24 2.2 Considerations and Sources of Uncertainty...... 28 2.2.1 Consideration/Source of Uncertainty 1: Unequal Diet-Tissue Fractionation Among Species...... 28 2.2.2 Consideration/Source of Uncertainty 2: Variable Diet-Tissue Fractionation Within a Species ...... 32 2.2.3 Consideration/Source of Uncertainty 3: Different Tissues From an Individual Animal Have Variable Stable Isotope Ratios...... 34 2.2.4 Consideration/Source of Uncertainty 4: Baseline Stable Isotope Ratios Vary Across Systems ...... 37 2.2.5 Consideration/Source of Uncertainty 5: The Presence of Omnivores..... 38 2.2.6 Consideration/Source of Uncertainty 6: Movement of Animals and Nutrients Between Food Webs/Ecosystems ...... 39 2.3 New Directions...... 41 2.4 Acknowledgements...... 45 2.5 References...... 45 2.6 Tables and Captions...... 66 2.7 Figures and Captions...... 68 Chapter 3.0 An Elemental and Stable Isotope Assessment of Water Strider Feeding Ecology and Lipid Dynamics: Synthesis of Lab and Field Studies...... 70 Abstract...... 70 3.1 Introduction...... 71 3.2 Methods...... 75 3.3 Results...... 80 3.4 Discussion...... 83

3.5 Acknowledgements...... 90 3.6 References...... 90 3.7 Tables and Captions...... 99 3.8 Figures and Captions...... 102 Chapter 4.0 Factors Affecting Water Strider (Hemiptera: Gerridae) Mercury Concentrations in Lotic Systems ...... 107 Abstract...... 107 4.1 Introduction...... 108 4.2 Methods...... 111 4.3 Results...... 118 4.4 Discussion...... 123 4.5 Acknowledgements...... 133 4.6 References...... 134 4.7 Tables and Captions...... 144 4.8 Figures and captions...... 146 Chapter 5.0 Mercury Biomagnification in Streams: The Role of Water Chemistry, Food Web Characteristics, and Proximity to an Emission Source ...... 154 Abstract...... 154 5.1 Introduction...... 155 5.2 Methods...... 159 5.3 Results...... 163 5.4 Discussion...... 168 5.5 Acknowledgements...... 176 5.6 References...... 177 5.7 Tables and Captions...... 189 5.8 Figures and Captions...... 194 Chapter 6.0 Conclusions and Recommendations ...... 199 6.1 Hg in Running Waters ...... 200 6.2 Merits of Different Sentinels ...... 202 6.3 Coal-fired Power Plants ...... 205 6.4 Mercury in New Brunswick, Canada...... 206 6.5 References...... 209 6.6 Tables and Captions...... 213 6.7 Figures and Captions...... 214 APPENDIX 1...... 216 APPENDIX 2...... 220 APPENDIX 3...... 222

Vita

vii

List of Tables

Table 2.1 Possible choices for diet-tissue 15N fractionation (∆15N) for carnivorous fishes when constructing food webs without a priori knowledge of ∆15N from captive studies ...... 66 Table 2.2 Features of different tissues used for stable isotope analysis in ecotoxicology studies ...... 67 13 15 Table 3.1 Stable isotope ratios (δ Cadj and δ N, in ‰) and C:N for A. remigis females (F), males (M) and nymphs (N) collected over the course of the growing season in 2007 in four streams in New Brunswick, Canada. Each point represents a single pooled sample. Table continued on next page...... 99 Table 4.1 Total Hg in water striders collected in eight recreational fishing areas (as designated by the provincial authority, see Fig. 1) in 2004 and two regions with point sources of Hg (Belledune and Grand Lake) in 2005 (n = # of sites sampled). Different superscripts indicate significantly different means...... 144 Table 4.2 Best-fit equations relating log-transformed water strider Hg concentrations and log-transformed water quality characteristics in New Brunswick, Canada streams. All equations are in the form y = mx + b...... 145 Table 5.1 Mercury biomagnification slopes measured in different ecosystems. TL = trophic level, FWMF = Food Web Magnification Factor, Type = methyl Hg (M), total Hg (T), or unknown (U)...... 189 Table 5.2 Multiple regression output with variables predicting Hg in individual blacknose dace (A) and brook trout (B) from New Brunswick streams...... 191 Table 5.3 Final model selections based on stepwise linear regression relating average Hg concentrations in blacknose dace (A) and brook trout (B) to explanatory variables (stream pH, distance from a coal fired power plant, sulfate, total organic carbon, total phosphorus and trophic level) in New Brunswick, Canada streams...... 192 Table 5.4 Correlations between baseline Hg concentrations (Hg at trophic level = 2) and Hg biomagnification slopes with selected environmental variables...... 193 Table 6.1 The number of brook trout and blacknose dace captured in two regions of New Brunswick, Canada with Hg concentrations above guidelines as set out by Health Canada (human consumption guideline of 0.5 ug g-1 wet weight) and Environment Canada (tissue residue guideline for fish eating wildlife of 0.06 ug g-1 wet weight)...... 213

viii

List of Figures

Figure 2.1 Theoretical biomagnification of contaminants in contrasting food webs with similar baseline and different slopes (A) and similar slopes and different baselines (B). Animals deriving greater than 70% of their biomass from carbon source 1 or carbon source 2 were separated using δ13C...... 68 Figure 2.2 Hypothetical consumer stable carbon (δ13C) and nitrogen (δ15N) isotopic ratios (solid circles), along with three (A) and four (B) potential food sources, and back- calculated dietary mixtures after correcting for diet-tissue 13C and 15N fractionation after Vander Zanden and Rasmussen (2001, +’s), Post (2002, open diamonds), and McCutchan et al. (2003, x’s). Note that in figure A, only when using fractionation estimates from McCutchan et al. (2003) are feasible source contributions obtained, while in scenario B, using the same fractionation estimates places the mixture outside of feasible source contributions...... 69 Figure 3.1 Wet weight (A, in mg, mean of multiple individuals) and C/N (B, individuals) of Aquarius remigis nymphs (circles), adult females (squares) and adult males (diamonds) vs. time in a laboratory diet-switch experiment. Animals were captured from a stream on day 0 and placed on a diet of crickets. Derived growth equations for female (F, n = 14) and male (M, n = 13) striders are shown in (A)...... 102 13 13 15 Figure 3.2 δ C (A, in ‰), δ Cadj (B, in ‰), and δ N (C, in ‰) in individual Aquarius remigis nymphs (circles), adult females (squares) and adult males (diamonds) vs. time in a laboratory diet-switch experiment (n = 52 for all three panels). Animals were captured from a stream on day 0 and placed on a diet of crickets (dashed line indicates isotope ratio of the diet). Open symbols = tank 1 (n = 2/date), solid symbols = tank 2 (n = 2/date), solid triangle = mean value (± S.D.) for a sample collected from the source stream on day 51...... 103 Figure 3.3 Percent aquatic carbon (mean values as estimated from mixing models using stable carbon isotopes) in the diet of Aquarius remigis (solid symbols) and Metrobates hesperius (open symbols) vs. stream order in New Brunswick, Canada streams. The best fit equation for A. remigis (n = 41) is included...... 105 Figure 3.4 Percent aquatic carbon (mean values as estimated from mixing models using stable carbon isotopes) in the diet of Aquarius remigis (solid symbols) and Metrobates hesperius (open symbols) vs. sampling date in Parks Brook, New Brunswick, Canada. Squares = females, diamonds = males, circles = nymphs. ... 106 Figure 4.1 Location of streams sampled in New Brunswick Canada in (A) 2004 (open circles) and 2005 (solid circles) and (B) 2006 and 2007. Point sources of Hg are marked with stars in the Belledune region and the Grand Lake region (A), and the Grand Lake power plant sits at the centre of the bullseye in 2006/2007 (B)...... 146 Figure 4.2 Mean total Hg concentrations (ug.g-1 d.w.) (A) and mean % sulphur (d.w.) (B) in Old Man’s Beard (Usnea sp.) relative to distance from a coal-fired power plant in New Brunswick, Canada in 2006 (open circles, solid best-fit line) and 2007 (solid diamonds, hatched best-fit line)...... 147 Figure 4.3 Mercury concentrations in female (solid diamonds, solid best-fit line) and male (open circles, hatched best-fit line) water striders (Aquarius remigis) in New Brunswick, Canada in 2006 (A) and 2007 (B) relative to distance from a coal-fired power plant (Fig. 1)...... 148

Figure 4.4 Correlation between male and female wet weights (A) and Hg concentrations (B) for Aquarius remigis in New Brunswick Canada streams in 2006 (open diamonds, solid best-fit line) and 2007 (solid diamonds, hatched best-fit line)..... 149 Figure 4.5 Correlation between male and female wet weights (A) and Hg concentrations (B) for Metrobates hesperius in New Brunswick Canada streams in 2006 (x, solid best-fit line) and 2007 (+, hatched best-fit line)...... 150 Figure 4.6 Mercury concentrations in Aquarius remigis in New Brunswick Canada streams in relation to (A) the percentage of aquatic carbon in the diet, and (B) δ15N as an indicator of trophic level for those streams with water striders having 0% aquatic carbon in the diet...... 151 Figure 4.7 Total Hg concentrations (ug g-1 d.w.) in Aquarius remigis females (solid diamonds), males (open circles) and nymphs (shaded triangles) from four New Brunswick, Canada streams during the growing season in 2006 and 2007...... 153 Figure 5.1 Total Hg concentrations (ug g-1 d.w.) in individual blacknose dace (solid diamonds) and brook trout (open circles) vs. body size (fork length, in cm) (A) and mean total Hg concentrations in dace and trout at sites where both species were captured (B). All stream sites were located in New Brunswick, Canada, and the dashed line in (A) indicates Health Canada’s recommended safe consumption limit of 0.5 ug g-1 wet weight (converted to 2.5 ug g-1 dry weight assuming 80% moisture)...... 194 Figure 5.2 Partial residual plots resulting from multiple regressions (see Table 5.3) of (A) pH and distance from a coal fired power plant on blacknose dace total Hg concentrations and (B) distance from the power plant and trophic level (TL) on brook trout total Hg concentrations in New Brunswick, Canada streams. All other variables (dace: trophic level, total organic carbon, sulphate, total phosphorus; trout: pH, total organic carbon, sulphate, total phosphorus) were not included in the final model...... 195 Figure 5.3 Length frequency histograms for brook trout (A, open bars, n = 123) and blacknose dace (B, solid bars, n = 294) in New Brunswick, Canada streams. In (C), blacknose dace (n = 60) from streams with pH < 7 are shown...... 197 Figure 5.4 Stable carbon ratios (δ13C, in ‰) of fish (blacknose dace = solid diamonds, brook trout = open circles) relative to biofilm (A) and primary consumers (B) in New Brunswick, Canada streams. Each point represents an average value for a single site. Solid lines indicate best-fit regressions for dace...... 198 Figure 6.1 Correlation of Hg concentrations in female A. remigis water striders with those of two species of fish – blacknose dace (solid diamonds) and brook trout (open circles) – in New Brunswick, Canada streams. Each point represents an average value from a single stream. Statistical testing was done with Spearman correlations...... 214 Figure 6.2 Conceptual model of Hg cycling in a temperate stream affected by local Hg emissions. Boxes show ecosystem compartments and circles show factors determining Hg concentrations in those compartments. Strong links are indicated by solid lines and weak links by hatched lines...... 215

Chapter 1.0 Introduction to the Problem

1.1 Background

From the Roman use of lead piping to methylmercury poisoning in Japan, exposure to toxic elements has been causing negative human health effects for millennia.

Public support for the regulation and monitoring of contaminants by government agencies and their study by academics was likely popularized by the publication of

“Silent Spring” (Carson 1962), a book that chronicled the damaging effects of pesticides on wildlife. In the intervening years since Silent Spring, the modern environmental movement has gained momentum, with the creation of numerous government agencies

(e.g. United States Environmental Protection Agency, established 1970; Environment

Canada, established 1971), and non-government organizations (e.g. Greenpeace, founded

1971; David Suzuki Foundation, incorporated 1990) with an environmental focus. While the public continues to support environmental legislation, criticism has been directed at the most vocal environmental advocates for allegedly distorting facts regarding environmental issues and misleading the public (Lomborg 2001). This leads to a tug of war between those who claim that money and attention spent on certain environmental issues is money and attention wasted (Lomborg 2001) and those who present somewhat of a doomsday scenario and urge that immediate action is warranted (Conkin 2007). The task of scientists who study environmental issues is to provide crucial, unbiased information regarding real environmental threats to the health (both short and long term) of humans and ecosystems to ensure that issues receive the appropriate amount of study and response. One such environmental issue is mercury (Hg), a contaminant that has

been in the crosshairs of the environment movement for over 50 years (Mergler et al.

2007).

Mercury is a naturally-occurring but widespread and toxic element. The potential dangers of Hg contamination of foods first received attention in the late 1950’s when residents of the Minamata Bay area of Japan developed severe neurotoxicological symptoms (Kudo et al. 1998). The effects were linked with the consumption of local fish that had been contaminated by methylmercury (methyl Hg) from waste dumping by a nearby chemical manufacturing company (Harada 1995). The publicity associated with the tragedy and those elsewhere (e.g. Troyer 1977) led several jurisdictions, including

New Brunswick, Canada (the location of the current study) to determine if resident fish were safe to eat (Zitko et al. 1971). It also stimulated a generation of research into the factors that contribute to high Hg concentrations in biota (Downs et al. 1998, Boening

2000, Wang et al. 2004), and to date it has proven difficult to gain a complete understanding of Hg as a global pollutant due its complex nature (Blais 2005).

Human activities have contributed to a profound shift in the global Hg cycle.

Since the dawn of the industrial revolution and the explosion of the human population,

Hg has been extracted from subsurface layers, concentrated, and broadcast to all corners of the globe via combustion processes (Schuster et al. 2002). Contamination has been particularly pronounced in aquatic systems, where sulfate-reducing bacteria may play a role in converting Hg from its elemental form to the methylated form that accumulates up food chains (Watras et al. 1998) and is toxic to humans and wildlife. As a result, numerous fish advisories are issued by government agencies with warnings to avoid consumption of fishes of a particular species or size (Cunningham et al. 1994). For

humans and wildlife, the primary route of exposure to contaminants such as Hg is via the

diet (Heinz and Hoffman 1998, Booth and Zeller 2005, Zahir et al. 2005). The source of food for wild animal populations is therefore important in determining their Hg concentrations.

Much of our understanding of Hg in aquatic systems stems from studies on its behaviour in lakes and wetlands (e.g. Snodgrass et al. 2000, Jeremiason et al. 2006,

Orihel et al. 2007). For example, it is now known that Hg concentrations in abiotic and

biotic compartments are affected by water chemistry, particularly pH, sulphate, and

dissolved organic carbon, with a variety of sample matrices (e.g. water, invertebrates,

fish, and birds) showing higher Hg concentrations in acidic, brown water systems than

those in non-acidic, clearwater systems (Watras et al. 1998, Greenfield et al. 2001,

Rennie et al. 2005). These problems are exacerbated by acid-rain deposition, which has

led to an increase in the number of waterbodies with low pH and greater rates of Hg methylation (Gilmour and Henry 1991). It is also known that in aquatic systems methyl

Hg biomagnifies (defined as an increase in concentration from diet to consumer).

Biomagnification of methyl Hg causes increases in concentration several orders of

magnitude from primary producers through primary, secondary and tertiary consumers;

as a result lakes with longer food chains have top predators with higher Hg

concentrations (Cabana et al. 1994, Cabana and Rasmussen 1994). Finally, while remote

lakes with no nearby Hg emission sources can contain fishes with high Hg concentrations

(Bodaly et al. 1993) and adjacent lakes with similar atmospheric Hg loading can have

vastly different Hg concentrations in the same species (Blais et al. 2006), links have

recently been found between Hg in precipitation and Hg in insects and fish at the

continental scale (Hammerschmidt and Fitzgerald 2005, 2006). Given this knowledge, the remaining unknowns regarding Hg are largely centred on the relative importance of each of these three factors (water chemistry, food chain properties, and atmospheric deposition) in governing Hg concentrations in aquatic consumers. Furthermore, despite our knowledge of Hg patterns in lakes and oceans, comparatively little is known about

Hg biomagnification in running waters.

In order to advance our understanding of the Hg cycle, we are turning to novel approaches. One such approach is stable isotope analysis (SIA). SIA takes advantage of naturally-occurring ratios of elements such as carbon (13C/12C), nitrogen (15N/14N), hydrogen (2H/1H, hereafter referred to as D/H for deuterium/hydrogen), and sulphur

(34S/32S) to trace the flow of organic matter through food webs. SIA has been used in the past to link animals and their Hg burdens to dietary sources in lakes and marine systems

(Kidd et al. 1995, Atwell et al. 1998, Nisbet et al. 2002). At this point in the dissertation, introduction to SIA as a research tool will be limited, given the extent of coverage provided on the topic in Chapter 2, a review paper entitled “Applications, considerations and sources of uncertainty when using stable isotope analysis in ecotoxicology” (Jardine et al. 2006). Measuring contaminant biomagnification in streams has rarely been attempted using SIA (Kidd 1998, Berglund et al. 2005) and, to my knowledge, no studies have been published that have measured Hg biomagnification in streams with SIA.

A second novel approach to the study of contaminants is in the use of sentinel species (Beeby 2001). Sentinel species provide ecologically-relevant information on sources, concentrations, or effects of various stressors, including contaminants, and may include insects (e.g. Jardine et al. 2005), mollusks (e.g. de Freitas Rebelo et al. 2003),

fishes (e.g. Munkittrick and Dixon 1989), mammals (e.g. Yates et al. 2005), and birds

(e.g. Fox et al. 2002). Ideal characteristics of sentinels for contaminant research include

ubiquity and abundance, known mobility, a predictable relationship between tissue and

source concentrations, and a well-known life history (Beeby 2001). The major purpose

in using sentinels is to gain information on spatial patterns of contaminants that would

otherwise be lacking due to reduced occurrence of target species (i.e. recreational fishes)

across sites or conservation concerns regarding their use in a sacrificial manner. The

three main sentinel species I will be using in this dissertation are a predatory insect (water striders, Hemiptera: Gerridae), a minnow (blacknose dace, Rhinichthys atratulus), and a

lichen (Old Man’s Beard, Usnea spp.). My use of these sentinel species will be limited to

indicators of Hg concentrations in the environment. In other words, I will not be

examining effects on the sentinels themselves; rather they will be used as “accumulator

species” (sensu Beeby 2001) to assess the relative magnitude of Hg contamination at a

given site.

The overarching goal of this dissertation is to understand the relative importance

of a variety of factors in enhancing biomagnification of Hg in stream ecosystems. These

factors include water chemistry, point source emissions, and food web attributes. These

three factors will be specifically addressed in Chapter 5. Prior to testing these factors,

however, I will initially introduce the concept of using SIA in ecotoxicological studies such as the current one (Chapter 2). The objective is to provide a sound understanding of the fundamentals of SIA in biomagnification studies and illustrate some of the challenges associated with using the approach.

Secondly, I will evaluate the possibility of using SIA and emerging sentinel

species as indicators of Hg contamination (Chapters 3 & 4). Particular attention will be paid to the water striders, which I have previously suggested may serve as useful surrogates for Hg concentrations in small fishes (Jardine et al. 2005) thus enabling more rapid and convenient determination of large scale patterns in Hg concentrations.

Attributes of the water strider, with a focus on the most common stream species Aquarius remigis, are examined in Chapters 3 and 4. Chapter 3 is an assessment of the feeding ecology of water striders using SIA and addresses the questions: What do they eat? How fast do they grow? How fast does carbon and nitrogen in their body tissues turn over?

These are important questions when trying to understand how these organisms acquire

Hg and how long it may remain in their tissues. Chapter 4 evaluates factors, some of which may prove confounding, that influence water strider Hg concentrations. It also examines more fully the life cycle of A. remigis from four index sites that were measured every two weeks for two years. I felt that both of these chapters were critical in determining whether the water strider was indeed an appropriate Hg sentinel in small fluvial systems (Jardine et al. 2005). The working hypothesis is that strider diet will be composed equally of aquatic and terrestrial carbon, and strider Hg concentrations will be negatively related to pH, positively related to total organic carbon content, increase with increasing atmospheric Hg deposition, and relate linearly to concentrations in stream fishes.

In Chapter 5, I take advantage of a unique situation in New Brunswick, Canada to assess the relative importance of the three above-mentioned factors (water chemistry, point source emissions, food web attributes) in determining Hg concentrations in stream

organisms. This unique situation is due to the presence of a coal-fired power plant in central New Brunswick at Newcastle Creek near Minto. This power plant burns low grade coal and, as a result, is a source of 100 kg of Hg annually to the surrounding environment (Environment Canada 2005). According to the National Pollutant Release

Inventory, it is the only major point source of Hg within a 200 km radius (Environment

Canada 2005), making it possible to attribute any local effects to its operation. For four years (2004-2007) I sampled biota in New Brunswick streams at varying distances from the plant. In the latter two years, sampling was conducted specifically to determine the influence of the plant. I used a bullseye design that is common in marine oil and gas monitoring (Green 2005) to determine the distance and direction of the effect. These measurements were intended to fill a gap in our understanding of Hg contamination in

New Brunswick. Previous work in the province found Hg concentrations for yellow perch (Perca flavescens) and brook trout (Salvelinus fontinalis) were similar to those found in adjacent provinces and states, but these studies focused on lakes in the north- central and south-west with little sampling occurring near Grand Lake (Kamman et al.

2005). For this study, it was hypothesized that Hg concentrations in fishes would be strongly related to water chemistry with additional variation explained by body size, trophic level and distance from the power plant. Larger fishes such as brook trout were expected to have higher Hg concentrations than minnows such as dace. Concentrations in fish muscle near Grand Lake were expected to be higher than average with the remainder of sites in New Brunswick expected to have fish with concentrations within the range observed for other locations. Increases in Hg concentrations with increasing trophic level in the food web were expected to be higher than those in other biomes (e.g.

marine systems) due to generally lower productivity in streams and slower growth rates

of consumers (Gross et al. 1988, Karimi et al. 2007). Concentrations in top predators

were expected to be controlled by both concentrations at the base of the food web and

subsequent biomagnification through the food web.

In the general discussion chapter (Chapter 6), I will attempt to synthesize the

information from the three experimental chapters and provide recommendations for

future research on Hg generally, and for running waters specifically. I will outline the

individual merits of the different sentinel species examined, provide some guidance as to

prediction of Hg concentrations in streams, and evaluate the likelihood that coal-fired

power plants have local effects on Hg concentrations in biota. I believe these are

important steps towards an integrated understanding of our potential impacts on the

natural environment.

Note on publication of articles and co-authors

This dissertation follows an articles format. Chapter 2 has been published in

Environmental Science and Technology. It originated as a conference presentation by

A.T. Fisk at the Society of Environmental Toxicology and Chemistry in Baltimore, 2005

(on which I was a co-author). In collaboration with co-authors K.A. Kidd and A.T. Fisk,

I later expanded the key points of the presentation into a review of the current knowledge

of SIA in ecotoxicology and outlined some of the major assumptions and directions for

future research. Chapter 3 has been published in Freshwater Biology. The study resulted from questions arising after my initial work on striders and Hg (Jardine et al. 2005) and summarizes four years of research into their feeding ecology on which I was the principal

investigator. I selected the sites, executed the field collections, processed the samples in

the laboratory, and analyzed the data. Co-authors K.A. Kidd and R.A. Cunjak provided

the conceptual framework for the investigation, and J.T. Polhemus provided the needed

resources for identifying species, life stages and sexes of striders. Chapter 4 has been

accepted with major revisions in Environmental Toxiclogy and Chemistry. It essentially

follows the same timeline as Chapter 3, with Hg data examined for sites where previous

collections were made. Again, K.A. Kidd and R.A. Cunjak provided the conceptual

framework and P.A. Arp provided technical expertise and the stimulus for the study as

part of the Collaborative Mercury Research Network. Chapter 5 has been submitted for

publication in Environmental Science and Technology. It captures data for two years of

full food web sampling (algae, invertebrates and fish) to understand Hg biomagnification

in streams. It has a single co-author (K.A. Kidd) and was completed by me in collaboration with her.

1.2 References

Atwell, L., Hobson, K.A., and Welch, H.E. 1998. Biomagnification and bioaccumulation

of mercury in an arctic marine food web: insights from stable nitrogen isotope

analysis. Canadian Journal of Fisheries and Aquatic Sciences 55: 1114-1121.

Beeby, A. 2001. What do sentinels stand for? Environmental Pollution 112: 285-298.

Berglund, O., Nystrom, P., and Larsson, P. 2005. Persistent organic pollutants in river

food webs: influence of trophic position and degree of heterotrophy. Canadian

Journal of Fisheries and Aquatic Sciences 62: 2021-2032.

Blais, J.M. 2005. Biogeochemistry of persistent bioaccumulative toxicants: processes

affecting the transport of contaminants to remote areas. Canadian Journal of

Fisheries and Aquatic Sciences 62: 236-243.

Blais, J.M., Charpentie, S., Pick, F., Kimpe, L.E., St.Amand, A., and Regnault-Roger, C.

2006. Mercury, polybrominated diphenyl ether, organochlorine pesticide, and

polychlorinated biphenyl concentrations in fish from lakes along an elevation

transect in the French Pyrenees. Ecotoxicology and Environmental Safety 63: 91-

99.

Bodaly, R.A., Rudd, J.W.M., Fudge, R.J.P., and Kelly, C.A. 1993. Mercury

concentrations in fish related to size of remote Canadian Shield lakes. Canadian

Journal of Fisheries and Aquatic Sciences 50: 980-987.

Boening, D.W. 2000. Ecological effects, transport, and fate of mercury: a general review.

Chemosphere 40: 1335-1351.

Booth, S., and Zeller, D. 2005. Mercury, food webs, and marine mammals: Implications

of diet and climate change for human health. Environmental Health Perspectives

113: 521-526.

Cabana, G., and Rasmussen, J.B. 1994. Modelling food chain structure and contaminant

bioaccumulation using stable nitrogen isotopes. Nature 372: 255-257.

Cabana, G., Tremblay, A., Kalff, J., and Rasmussen, J.B. 1994. Pelagic food chain

structure in Ontario lakes: a determinant of mercury levels in lake trout (Salvelinus

namaycush). Canadian Journal of Fisheries and Aquatic Sciences 51: 381-389.

Carson, R. 1962. Silent Spring. Houghton Mifflin Company, Boston, MA, USA. 400 pp.

Conkin, P.K. 2007. The State of the Earth. The University Press of Kentucky, Lexington,

KY, 308 pp.

Cunningham, P.A., Smith, S.L., Tippett, J.P., and Greene, A. 1994. A national fish

consumption advisory database – a step towards consistency. Fisheries 19: 14-23. de Freitas Rebelo, M., Reboucas do Amaral, M.C., and Pfeiffer, W.C. 2003. High Zn and

Cd accumulation in the oyster Crassostrea rhizophorae, and its relevance as a

sentinel species. Marine Pollution Bulletin 46: 1341-1358.

Downs, S.G., MacLeod, C.L., and Lester, J.N. 1998. Mercury in precipitation and its

relation to bioaccumulation in fish: a literature review. Water, Air, and Soil

Pollution 108: 149-187.

Environment Canada. 2005. National Pollutant Release Inventory.

www.ec.gc.ca/pdb/npri.

Fox, G.A., Grasman, K.A., Hobson, K.A., Williams, K., Jeffrey, D., and Hanbridge, B.

2002. Contaminant residues in tissues of adult and prefledged herring gulls from the

Great Lakes in relation to diet in the early 1990s. Journal of Great Lakes Research

28: 643-663.

Gilmour, C.C., and Henry, E.A. 1991. Mercury methylation in aquatic systems affected

by acid deposition. Environmental Pollution 71: 131-169

Green, R.H. 2005. Marine coastal monitoring: Designing an effective offshore oil and gas

environmental effects monitoring program. In: Armsworthy, S.L., Cranford, P.J.,

Lee, K. (Eds). Offshore oil and gas environmental effects monitoring – Approaches

and Technologies. Batelle Press, pp. 373-398.

Greenfield, B.K., Hrabik, T.R., Harvey, C.J., and Carpenter, S.R. 2001. Predicting

mercury levels in yellow perch: use of water chemistry, trophic ecology, and spatial

traits. Canadian Journal of Fisheries and Aquatic Sciences 58: 1419-1429.

Gross, M.R., Coleman, R.M., and McDowall, R.M. 1988. Aquatic productivity and the

evolution of diadromous fish migration. Science 239: 1291-1293.

Hammerschmidt, C.R., and Fitzgerald, W.F. 2005. Methylmercury in mosquitoes related

to atmospheric mercury deposition and contamination. Environmental Science and

Technology 39: 3034-3039.

Hammerschmidt, C.R., and Fitzgerald, W.F. 2006. Methylmercury in freshwater fish

linked to atmospheric mercury deposition. Environmental Science and Technology

40: 7764-7770.

Harada, M. 1995. Minamata Disease – Methylmercury poisoning in Japan caused by

environmental pollution. Critical Reviews in Toxicology 1: 1-24.

Heinz, G.H., and Hoffman, D.J. 1998. Methylmercury chloride and selenomethionine

interactions on health and reproduction in mallards. Environmental Toxicology and

Chemistry 17: 139-145.

Jardine, T.D., Al, T.A., Macquarrie, K.T.B., Ritchie, C.D., Arp, P.A., Maprani, A., and

Cunjak, R.A. 2005. Water striders (family Gerridae): Mercury sentinels in small

freshwater ecosystems. Environmental Pollution 134: 165-171.

Jardine, T.D., Kidd, K.A., and Fisk, A.T. 2006. Applications, considerations and sources

of uncertainty when using stable isotope analysis in ecotoxicology. Environmental

Science and Technology 40: 7501-7511.

Jeremiason, J.D., Engstrom, D.R., Swain, E.B., Nater, E.A., Johnson, B.M.,

Almendinger, J.E., Monson, B.A., and Kolka, R.K. 2006. Sulfate addition increases

methylmercury production in an experimental wetland. Environmental Science and

Technology 40: 3800-3806.

Kamman, N.C., Burgess, N.M., Driscoll, C.T., Simonin, H.A., Goodale, W., Linehan, J.,

Estabrook, R., Hutcheson, M., Major, A., Scheuhammer, A.M., and Scruton, D.A.

2005. Mercury in freshwater fish of northeast North America – A geographic

perspective based on fish tissue monitoring databases. Ecotoxicology 14: 163-180.

Karimi, R., Chen, C.Y., Pickhardt, P.C., Fisher, N.S., and Folt, C.L. 2007. Stoichiometric

controls of mercury dilution by growth. Proceedings of the National Academy of

Sciences U.S.A. 104: 7477-7482.

Kidd, K.A. 1998. Use of stable isotope ratios in freshwater and marine biomagnification

studies. In Environmental Toxicology: Current Developments. Edited by J. Rose.

Gordon & Breach Science Publishers, Amsterdam pp. 357-376.

Kidd, K.A., Hesslein, R.H., Fudge, R.J.P., and Hallard, K.A. 1995. The influence of

trophic level as measured by δ15N on mercury concentrations in freshwater

organisms. Water, Air, and Soil Pollution 80: 1011-1015.

Kudo, A., Fujikawa, Y., Miyahara, S., Zheng, J., Takigami, H., Sugahara, M., and

Muramatsu, T. 1998. Lessons from Minamata mercury pollution, Japan – After a

continuous 22 years of observation. Water Science and Technology 38: 187-193.

Lomborg, B. 2001. The Skeptical Environmentalist. Cambridge University Press,

Cambridge, UK, 515 pp.

Mergler, D., Andersen, H., Chan, H.M., Mahaffey, K., Murray, M., Sakamoto, M., and

Stern, A. 2007. Methylmercury exposure and health effects in humans: a worldwide

concern. Ambio 36: 3-11.

Munkittrick, K.R., and Dixon, D.G. 1989. Use of white sucker (Catostomus commersoni)

populations to assess the health of aquatic ecosystems exposed to low-level

contaminant stress. Canadian Journal of Fisheries and Aquatic Sciences 46: 1455-

1462.

Nisbet, I.C.T., Montoya, J.P., Burger, J., and Hatch, J.J. 2002. Use of stable isotopes to

investigate individual differences in diets and mercury exposures among common

terns Sterna hirundo in breeding and wintering grounds. Marine Ecology Progress

Series 242: 267-274.

Orihel, D.M., Paterson, M.J., Blanchfield, P.J., Bodaly, R.A., and Hintelmann, H. 2007.

Experimental evidence of a linear relationship between inorganic mercury loading

and methylmercury accumulation by aquatic biota. Environmental Science and

Technology 41: 4952-4958.

Rennie, M.D., Collins, N.C., Purchase, C.F., and Tremblay, A. 2005. Predictive models

of benthic invertebrate methylmercury in Ontario and Quebec lakes. Canadian

Journal of Fisheries and Aquatic Sciences 62: 2770-2783.

Schuster, P.F., Krabbenhoft, D.P., Naftz, D.L., Cecil, L.D., Olson, M.L., Dewild, J.F.,

Susong, D.D., Green, J.R., and Abbott, M.L. 2002. Atmospheric mercury

deposition during the last 270 years: A glacial ice core record of natural and

anthropogenic sources. Environmental Science and Technology 36: 2303-2310.

Snodgrass, J.W., Jagoe, C.H., Bryan, A.L. Jr., Brant, H.A., and Burger, J. 2000. Effects

of trophic status and wetland morphology, hydroperiod, and water chemistry on

mercury concentrations in fish. Canadian Journal of Fisheries and Aquatic Sciences

57: 171-180.

Troyer, W. 1977. No Safe Place. Clarke, Irwin & Company. Toronto, Canada. 267 pp.

Wang, Q.R., Kim, D., Dionysiou, D.D., Sorial, G.A., and Timberlake, D. 2004. Sources

and remediation for mercury contamination in aquatic systems – a literature review.

Environmental Pollution 131: 323-336.

Watras, C.J., Back, R.C., Halvorsen, S., Hudson, R.J.M., Morrison, K.A., and Wente,

S.P. 1998. Bioaccumulation of mercury in pelagic freshwater food webs. The

Science of the Total Environment 219: 183-208.

Yates, D.E., Mayack, D.T., Munney, K., Evers, D.C., Major, A., Kaur, T., and Taylor,

R.J. 2005. Mercury levels in mink (Mustela vison) and river otter (Lontra

canadensis) from northeastern North America. Ecotoxicology 14: 263-274.

Zahir, F., Rizwi, S.J., Haq, S.K., and Khan, R.H. 2005. Low dose mercury toxicity and

human health. Environmental Toxicology and Pharmacology 20: 351-360.

Zitko, V., Finlayson, B.J., Wildish, D.J., Anderson, J.M., and Kohler, A.C. 1971.

Methylmercury in freshwater and marine fishes in New Brunswick, in the Bay of

Fundy, and on the Nova Scotia banks. Journal of the Fisheries Research Board of

Canada 28: 1285-1291.

Chapter 2.0 Applications, Considerations, and Sources of Uncertainty When Using

Stable Isotope Analysis in Ecotoxicology

Timothy D. Jardine, Karen A. Kidd, and Aaron T. Fisk

Published in Environmental Science and Technology 2006, Vol. 40: 7501-7512.

Abstract

Stable isotope analysis (SIA) has become a powerful tool for ecotoxicologists to study dietary exposure and biomagnification of contaminants in wild animal populations. The use of SIA in ecotoxicology continues to expand and, while much more is known about the mechanisms driving patterns of isotopic ratios in consumers, there remain several considerations or sources of uncertainty that can influence interpretation of data from field studies. Current uses of SIA in ecotoxicology are outlined, including estimating the importance of dietary sources of carbon and their application in biomagnification studies, and six main considerations or sources of uncertainty associated with the approach are presented: 1) unequal diet-tissue stable isotope fractionation among species, 2) variable diet-tissue stable isotope fractionation within a given species, 3) different stable isotope ratios in different tissues of the animal, 4) fluctuating baseline stable isotope ratios across systems, 5) the presence of true omnivores, and 6) movement of animals and nutrients between food webs. Since these considerations or sources of uncertainty are difficult to assess in field studies, researchers are urged to consider the following in designing ecotoxicological research and interpreting results: assess and utilize variation in stable isotope diet-tissue fractionation among animal groups available in the literature;

determine stable isotope ratios in multiple tissues to provide a temporal assessment of feeding; adequately characterize baseline isotope ratios; utilize stomach contents when possible; and assess and integrate life history of study animals in a system.

2.1 Introduction

Naturally occurring stable isotope ratios of carbon (13C/12C or δ13C), nitrogen

(15N/14N or δ15N), sulphur (34S/32S or δ34S), hydrogen (2H/1H or D/H or δD) and oxygen

(18O/16O or δ18O) have become powerful and popular tools to trace organic matter

sources in ecosystems. Isotopic ratios are calculated following the formula:

δX = (Rsample/Rstandard - 1) * 1000

13 where: X is the heavier isotope (e.g. C), Rsample is the raw ratio of the heavy to light

isotope in the sample, and Rstandard is the raw ratio of the heavy to light isotope in an

internationally accepted standard. These standards include variations of: Peedee

belemnite (PDB) carbonate for δ13C, Atmospheric Nitrogen (AIR) for δ15N, Canyon

Diablo troilite (CDT) for δ34S, and Standard Mean Ocean Water (SMOW) for δD and

δ18O (Werner and Brand 2001)

The use of stable isotope analysis (SIA) in studies of animal ecology rests on two

central tenets: 1) stable isotope ratios in consumers are proportional to those in their diet,

and 2) differences in isotope ratios exist among food sources available for consumers.

Early laboratory studies by DeNiro and Epstein (1978, 1981) established the basis of the

first tenet. They reported that ratios of lab-reared consumers had stable carbon ratios that were not statistically different than their food (mean difference = 0.8 ± 1.1‰ S.D.), while stable nitrogen ratios of consumers were enriched in nitrogen-15 relative to their diet

(mean difference = 3.0 ± 2.6‰ S.D.). These relationships have since been corroborated in a series of literature syntheses of lab-based trials (Vander Zanden and Rasmussen

2001, Post 2002, McCutchan et al. 2003). The second tenet has been demonstrated repeatedly in a variety of field studies that showed differences among different food types, including C3 vs. C4 plants (O’Leary 1988) and organic matter originating from aquatic vs. terrestrial (Rounick and Winterbourn 1986), marine vs. freshwater (Fry and

Sherr 1984) and planktonic vs. benthic (France 1995, Hecky and Hesslein 1995, Vander

Zanden and Rasmussen 1999) habitats.

These two central tenets form the foundation of stable isotope ecology that is practiced today. One isotope (typically carbon) shows considerable isotopic change during its fixation by primary producers, while another isotope (typically nitrogen) shows considerable change as it is processed by consumers. The combination of these two isotopes, therefore, allows the investigation of different energy flow processes that shape the structure and function of food webs. SIA has advantages over traditional dietary analyses (e.g. gut contents) because it provides a time-integrated representation of assimilated food rather than a snapshot of recently ingested items.

Coincident with the increase in knowledge concerning SIA has been an expansion in its applications. One such expansion is into the field of ecotoxicology, wherein SIA is used to examine variability in contaminant concentrations of animal populations.

Because of the importance of diet as a route of exposure for heavy metals,

organochlorines, and other persistent contaminants (Thomann and Connolly 1984, Hall et

al. 1997), SIA has considerably advanced the field of ecotoxicology by linking wild

animal populations to their diet and ultimate contaminant source. Ecotoxicology studies that use SIA can be considered in three categories: 1) qualitative linkages between dietary habits of animal populations and resultant contaminant concentrations, 2) food web biomagnification studies, and 3) quantitative assessments of habitat-specific foraging as a

means of explaining contaminant concentrations. The strongest studies combine

categories 2 and 3, simultaneously assessing the importance of both food chain length and underlying dietary pathway in determining contaminant concentrations in an organism. Estimating trophic position and determining underlying organic matter pathways is typically accomplished with δ15N and δ13C, respectively, while those

elements as well as δ34S, δD, and δ18O are increasingly being used to examine larger

scale animal movement patterns (Hobson 1999). This review will focus attention on how

δ13C and δ15N have been used in the past to relate contaminant contaminants in an

organism to its dietary characteristics, and will highlight six of the main

considerations/sources of uncertainty when using the technique.

2.1.1 Qualitative Studies

In aquatic systems, food as opposed to water is often the major exposure route of

animals to contaminants, particularly mercury (Hall et al. 1997) and hydrophobic organic

contaminants such as PCBs and DDT (Thomann and Connolly 1984). For air-breathing

organisms, food is usually the only exposure route for contaminants. Thus, researchers

who study contaminant fate in biota may wish to examine the diet of animal populations

and may use isotope data to generate hypotheses as to why certain subpopulations or

specific taxa have higher or lower contaminant concentrations than their counterparts

occupying similar habitats. In these studies, contaminant concentrations are not directly

or quantitatively linked to isotope ratios, rather isotope ratios are first compared

statistically among taxa, followed by the analyses of inter-taxa differences in contaminant

concentrations. For example, the scavenging gastropod Cyclope neritea had elevated

δ15N when compared to other organisms in the River Po delta in Italy, and its higher

trophic position was used to explain the gastropod’s higher concentrations of mercury,

cadmium, lead, copper, and zinc (Camusso et al. 1998). Other examples of these

statistical approaches include examinations of trace metals in marine food webs (Das et

al. 2000), organochlorines and metals in avian tissues (Mazak et al. 1997, Braune et al.

2001, 2002, Fox et al. 2002), and mercury in riverine fishes (Gustin et al. 2005). Other such studies have focused on a single species and used SIA to explain variability in contaminant concentrations within these species. They include analyses of mercury in great skua (Catharacta skua) chicks (Bearhop et al. 2000) and common terns (Sterna hirundo, Nisbet et al. 2002), organochlorines in walrus (Odobenus rosmarus, Muir et al.

1995), Atlantic salmon (Salmo salar, Berglund et al. 2001), glaucous gulls (Larus hyperboreus, Sagerup et al. 2002) and Arctic fox (Alopex lagopus, Hoekstra et al. 2003), and metals in raccoons (Procyon lotor, Gaines et al. 2002).

2.1.2 Biomagnification Studies

The studies described above, while valuable in providing some explanation for the variability observed in animal contaminant concentrations, do not offer insight into

ecosystem scale patterns, and typically use SIA in a semi-quantitative fashion. One of the advantages of SIA is its quantitative nature that has led to a movement away from categorical foraging data (Vander Zanden and Rasmussen 1996). As one such example, an initial study in eastern Canada had classified lakes based on the presence or absence of different forage species (e.g. mysids, forage fishes), assigned organisms to discrete trophic levels (TLs) and used this information to predict/describe concentrations of contaminants in top predators across systems (Rasmussen et al. 1990). Lakes with long food chains due to the presence of mysids and/or forage fishes had lake trout (Salvelinus namaycush) with higher contaminant concentrations than lakes with short food chains lacking intermediate trophic links (Cabana et al. 1994). Follow-up studies on these same systems used SIA and showed that omnivory was common in the fish species, resulting in some systems predicted to have lake trout with high Hg concentrations having lower concentrations than expected. This demonstrated that δ15N is a more powerful approach than discrete TL classifications in predicting contaminant concentrations of higher order predators (Cabana and Rasmussen 1994, 1996).

The first study that used SIA to quantify the trophic transfer or biomagnification of contaminants through entire food webs was Broman et al. (1992), who used regressions relating concentrations of polychlorinated dibenzo-p-dioxins (PCDDs) to

δ15N following:

[PCDD] = e(b + m*δ15N)

which transforms to:

ln [PCDD] = b + m*δ15N

In this equation, the slope (m) is a measure of the biomagnification of PCDD through the food web, whereas the intercept (b) may represent the concentration of the contaminant at the base of the food web, although this requires further investigation. Recently, the δ15N values in equation 3 have been replaced with TL estimates that are based on the δ15N value of an organism and are calculated using the equation (Hobson and Welch 1992):

15 15 TLconsumer = (δ Nconsumer – δ Nprimary producer) / ∆15N + 1

15 where δ Nprimary producer is assumed to occupy TL 1 and ∆15N (the “enrichment factor”)

represents the increase in δ15N from one TL to the next. Substituting TL for δ15N in the

earlier equation gives:

ln or log [contaminant] = b + m * TL

and a food web magnification factor [FWMF; has also been called a trophic

magnification factor (TMF)] can be calculated (Fisk et al. 2001a):

FWMF = em (or 10m)

The advantage of the FWMF over a δ15N-derived slope is that the former

represents the average increase in contaminant concentrations from one TL to the next,

and is analogous to a biomagnification factor for the food web (Borga et al. 2004). In

contrast, the slopes of the contaminant-δ15N regressions quantify the increase in

contaminant concentrations with each per mil δ15N, a measure that is much more abstract

with respect to its application to prey to predator transfer of these contaminants.

Furthermore, the use of TL allows unique enrichment factors for species or groups of

animals to be incorporated into the FWMF estimates of trophic transfer of contaminants.

More specifically, food web studies that have included birds with invertebrates, fish and mammals can utilize different enrichment factors for these different taxa (Fisk et al.

2001a, Hobson et al. 2002). Both unadjusted and adjusted (for different diet-tissue 15N enrichment) contaminant-TL relationships describe the average increase in the contaminant with TL and can therefore be useful in assessing biomagnification within and across systems.

Over the past decade SIA has been valuable in assessing the biomagnification potential of a variety of contaminants. While certain contaminants (and congeners) increase in concentration with increasing TL in certain ecosystems (e.g. organochlorines,

Fisk et al. 2001a, Jarman et al. 1996, Kidd et al. 1995a,b, 1998a,b, 2001, Kiriluk et al.

1995, Kucklick and Baker 1998; mercury, Jarman et al. 1996, Atwell et al. 1998, Kidd et al. 2003, Evans et al. 2005; and rubidium and cesium, Campbell et al. 2005a) others have shown no significant relationship or have decreased significantly when regressed against

δ15N of the biota within the food web (e.g. selenium, lead, cadmium and copper, Jarman

et al. 1996, Moisey et al. 2001, Quinn et al. 2003, Mackintosh et al. 2004, Campbell et al.

2005b). Assessing the potential for a particular contaminant to biomagnify in a

lacustrine, riverine or marine food web is important because those chemicals that show

limited biomagnification potential presumably pose less of a threat to the health of

humans and wildlife that rely on aquatic food webs as a source of food. As a result,

biomagnification potential has been advocated as a means for determining the level of

regulation (i.e. limitation of releases to the environment) to assign to different industrial chemicals in bureaucratic legislation (Mackay and Fraser 2000).

2.1.3 Quantitative Assessment of Carbon Sources

Another factor that may affect an organism’s exposure to and accumulation of contaminants is the habitat in which it forages relative to other individuals within a population. The first paper to demonstrate a link between foraging patterns and contaminants was Spies et al. (1989), who showed a relationship between a variety of contaminants in marine fishes and their carbon, nitrogen, and hydrogen isotope ratios that

were related to exposure to a sewage outfall. Fishes with less “heavy” carbon and nitrogen isotopic ratios derived a larger proportion of biomass from the sewage outfall, and a significant negative trend was observed between δ15N and organochlorines,

suggesting greater exposure in fish that were connected to the sewage outfall (Spies et al.

1989).

Mathematical formulas have been created to quantify source contributions to an

organism using stable isotopes, assuming the dietary sources differ enough in their

isotope ratios to allow separation (Fry and Sherr 1984). Calculation of the relative

importance of two dietary sources to an animal is achieved by quantifying the isotope

ratios and calculating the relative isotopic difference of the sources, and then comparing

the organism’s isotope ratio to one of these sources of food. In a simple two source, one

isotope mixing model, dietary carbon source is calculated according to (Dunton and

Schell 1987):

13 13 13 13 Pa = (δ Corganism-δ Cb)/(δ Ca-δ Cb)

13 where Pa is the proportion of source a (e.g. pelagic carbon) in the diet, and δ Ca and

13 δ Cb are the baseline ratios for source a and b respectively. In this model, there is no

fractionation of carbon-13 through the food web (e.g. Vander Zanden and Vadeboncoeur

2002), but if necessary this can be accounted for by adjusting source ratios by an

appropriate amount (e.g. 0.8‰, Vander Zanden and Rasmussen 2001).

One of the limitations associated with these mixing equations is their inability to reflect the error around estimates of source and mixture mean values, a problem that has been overcome by the development of mixing model theory and software that covers a variety of scenarios where mixing models are appropriate. These include: standard one- isotope, two source and two-isotope, three source models (Isoerror, Phillips and Gregg

2001); concentration dependent mixing models that account for the carbon and nitrogen stoichiometry of various food types (Isoconc, Phillips and Koch 2002) and; mixing models that provide probability distributions for various sources when the number of sources far outnumbers the number of isotopes available for analysis (Isosource, Phillips and Gregg 2003), such as the separation of littoral, pelagic, and profundal energy sources using only δ13C in lakes. These models also conserve mass balance, a feature that was

lacking in earlier models. It should also be noted that in some cases it may not be

possible or necessary to assess the relative importance of different carbon sources

because the food web is driven by energy from a single food source (e.g. phytoplankton,

Atwell et al. 1998).

Multiple stable isotopes can be used to understand and quantify the influence of

both underlying dietary pathways and trophic positioning on concentrations of contaminants in individuals and populations. For example, Kidd et al. (2001, 2003) used

δ13C to delineate organisms connected to the benthic and pelagic food chains in Lake

Malawi, Africa. The biomagnification slope for ∑DDT (vs δ15N) was higher in the

pelagic than the benthic food web, resulting in higher ∑DDT in top predators from the

former niche in the system (Kidd et al. 2001). In the same lake, concentrations of Hg

were also higher in the biota relying upon pelagic carbon despite similar biomagnification

slopes for that contaminant between the benthic and pelagic food webs; pelagic feeders

had higher Hg most likely due to higher inputs of Hg at the base of this food web (Kidd

et al. 2003).

Figure 2.1 illustrates the theoretical expectations for isotope and contaminant data

under these scenarios. Animals are assigned to one of two habitats based on δ13C (with

>70% contribution as a criteria, Kidd et al. 2003), and contaminant concentrations plotted

against δ15N to assess biomagnification. In Figure 2.1a, both habitats have similar baseline contaminant concentrations, but the slope is much steeper in habitat A. In

Figure 2.1b, the contaminant-TL slopes are the same but a higher baseline contaminant level in habitat A results in increased contaminant concentrations in all organisms from

that food web. As application of this technique increases, analyses of covariance (Sokal

and Rohlf 1995) can be used to assess if slopes or intercepts differ among two habitat

types within one system or among systems. Power et al. (2002) also demonstrated the

importance of both TL and carbon source in determining fish Hg concentrations in a sub-

Arctic lake. Similar to the Lake Malawi food web (Kidd et al. 2003), fishes that derived more energy from the pelagic zone had higher Hg than those connected to the benthic pathway (Power et al. 2002).

Separating food source pathways along with measures of TL has also been used to understand the fate of contaminants. Croteau et al. (2005) identified phytoplankton and epiphytes as two major primary food sources for consumers in the Sacramento-San

Joaquin River delta using stable carbon isotopes. They found a significant increase in

cadmium concentrations with increasing TL in the epiphyte food web that was not

apparent when the entire food web was considered collectively. Similarly, Berglund et

al. (2005) traced PCBs in Swedish streams and found higher concentrations in animals

associated with the detrital pathway compared to the algal pathway. PCB concentrations

also increased from primary producers through invertebrates to brown trout (Salmo

trutta), suggesting that both TL and carbon sources were important in those systems.

Testing the influence of both factors (carbon source and TL) also allows a comparison of

their relative importance. Campbell et al. (2000) found that lipid-rich animals connected

to the pelagic zone in Bow Lake, Canada had higher organochlorine concentrations than

those feeding in benthic habitats, and the effect of lipid and carbon source was far more

dominant than TL (estimated by δ15N) in explaining variability in biotic concentrations.

2.2 Considerations and Sources of Uncertainty

While SIA has emerged as a powerful tool to study a variety of ecological

applications, several considerations or sources of uncertainty remain when using the

technique. The following discussion will focus on those that are most relevant to the use

of SIA in ecotoxicology studies and some of their associated caveats. It will not discuss

analytical error, which has been covered previously (Jardine and Cunjak 2005).

2.2.1 Consideration/Source of Uncertainty 1: Unequal Diet-Tissue Fractionation

Among Species

In order to quantify the biomagnification of a contaminant through a food web,

the slope of the regression between, e.g., an organochlorine and the δ15N of the biota is

used as an overall descriptor and it is not necessary to assign a particular value to ∆15N

(the difference in δ15N between an animal and its diet). It must simply be agreed that it approaches some particular value when averaged over the entire food web and that the slopes of contaminant vs. δ15N regression reflects biomagnification and not changing

∆15N. However, when converting stable nitrogen ratios to TL estimates (Vander Zanden

and Rasmussen 1996) a value for ∆15N must be assigned. This is also necessary when deriving biomagnification factors (BMFs) within different compartments of the food web

(Fisk et al. 2001a) to compare results of those with prior studies that did not use SIA (e.g.

Russell et al. 1995). BMFs are calculated by dividing the ratio of predator to prey contaminant concentrations by the ratio of their TLs (Fisk et al. 2001a).

Recently, the use of primary consumers over primary producers as a baseline of a particular food web has led to the modification of equation 4. TLs of organisms in the food web are typically calculated according to:

15 15 TLorganism = (δ Norganism – δ Nbaseline)/∆15N + 2

15 15 where δ Nbaseline is the measured δ N of a long-lived primary consumer (TL = 2). A value between 2 and 5‰ is often assigned to ∆15N, although this may vary depending on the types of organisms within the food web and is the subject of some debate (Vanderklift and Ponsard 2003). Bivalve mollusks such as mussels (Cabana and Rasmussen 1996,

Vander Zanden and Rasmussen 1999, Post 2002) and clams (Fry 1999, Post 2002,

Swanson et al. 2003) have commonly been used to standardize the baseline of food webs, but other taxa have been used including gastropods (Kidd et al. 1998), copepods (Moisey et al. 2001, Fisk et al. 2001a, 2003, Campbell et al. 2005a), and other invertebrates

(Vander Zanden and Rasmusssen 1999). When baseline δ15N varies between two habitats within an ecosystem, the trophic position model accounts for individuals foraging in these habitat zones (e.g. pelagic vs. littoral zones of lakes) using:

15 15 15 TP = 2 + {δ Norganism – [δ Na * Pa + δ Nb * (1-Pa)]}/ ∆15N

15 15 where δ Na and δ Nb are stable nitrogen ratios of primary consumers from habitat a and b, respectively.

The choice of ∆15N (and even ∆13C) becomes important in these models because

it affects calculations of the importance of different sources to an organism’s diet. When using the overall mean ∆15N and ∆13C reported by multiple authors (Minagawa and

Wada 1984, Vander Zanden and Rasmussen 2001, Post 2002, McCutchan et al. 2003,

Vanderklift and Ponsard 2003) calculations result in the consumer’s mixture sometimes falling outside the mixing space used to estimate proportional contributions of different diet items (Figure 2.2, Phillips and Gregg 2001). This stems from spatial patterns in isotopic distributions and trophic fractionation, wherein δ13C and δ15N are often

correlated (with a positive r, i.e. foods that are enriched in 15N are often also enriched in

13C, and both 15N and 13C increase to some degree up the food chain); this, in turn,

generates obtuse triangles (Fig. 2a) or laterally compressed polygons (Fig. 2b) in dual

isotope space (Ben-David et al. 1997). This leaves little isotopic “space” for the mixture

to fall within and to provide feasible data on relative contributions of different sources to

an organism’s diet. In this case, researchers are left with the dilemma of choosing some

other value for ∆15N and ∆13C. For example, if the species of interest is northern pike

(Esox lucius), one could choose literature estimates for ∆15N of 3.4‰ (Minagawa and

Wada 1984, Post 2002), 2.9‰ (Vander Zanden and Rasmussen 2001), 2.5‰ (Vanderklift

and Ponsard 2003) or 2.0‰ (McCutchan et al. 2003) (Table 1). These are pooled

estimates for all animals, on all types of diet, and from a number of different

environments. Within each of these studies attempts have been made to account for

differences in ∆15N based on taxonomy, method of nitrogenous waste excretion, etc.,

with the hope that this could provide a more accurate value of ∆15N for the animal of

interest. However, the range of values for ∆15N in the literature still leaves a difficult

decision for the researcher studying northern pike, which is a carnivorous fish that excretes ammonia. It is not clear if the value of 2.7‰ for carnivores or 2.0‰ for animals that excrete ammonia should be used (Vanderklift and Ponsard 2003). The only true solution in this case is to rear northern pike on a known diet and then calculate ∆15N directly, but this could be a costly and labor-intensive exercise as it may take years to achieve isotopic equilibrium of the fish with its diet (MacNeil et al. 2006) and ∆15N still may be altered by local conditions.

Vander Zanden and Rasmussen (2001) argued that this assumption is less of a concern when using δ15N to measure biomagnification because variable values of ∆15N will be averaged out over multiple TLs and approach 3.4‰, especially when primary consumers are used as the baseline. Vander Zanden and Rasmussen (2001) showed far greater ∆15N variability in herbivores compared to carnivores, therefore by eliminating the highly variable initial link between primary producers and primary consumers by using shellfish or another obligate herbivore as a baseline in biomagnification studies

(Fisk et al. 2003), confidence in 3.4‰ as a value for ∆15N increases. As well, 3.4‰ as a value for ∆15N becomes more plausible with the inclusion of vertebrate species, which tend to be large and long-lived and less vulnerable to rapid isotopic turnover as compared to invertebrates. For example, strong relationships between contaminant concentrations and TL or δ15N were found for a food web that included zooplankton and vertebrates

(Fisk et al. 2001a), but when the contaminant-TL relationships only considered zooplankton the relationships were not as strong or were not significant (Fisk et al. 2003).

This may be due in part to issues such as body size and habitat that may be more

important in determining contaminant concentrations in zooplankton than their feeding

ecology (Borga et al. 2002a, b).

2.2.2 Consideration/Source of Uncertainty 2: Variable Diet-Tissue Fractionation

Within a Species

In the example above, if the researcher had done a captive rearing experiment to determine ∆15N and ∆13C for northern pike a further assumption is required before the results are applied to field studies. It must be assumed that the ∆15N values are static and not affected by environmental conditions such as temperature that would fluctuate over seasons for wild animals. Experimental evidence suggests that this assumption may be invalid. Changes in isotopic signatures of Daphnia magna on a fixed diet were related to differences in temperature across tanks (Power et al. 2003). Changes in walleye δ15N were independent of dietary switches or growth but were significantly related to age

(Overman and Parrish 2001), suggesting that animals become progressively enriched in

15N over time despite always feeding at the same TL. This could confound interpretation

of TL estimates for older consumers. However, no relationship between age and δ15N was found for lake trout (Kiriluk et al. 1995), and size, as a surrogate for TL, is a better predictor of δ15N than age among rainbow smelt (Osmerus mordax) morphotypes

(Jardine and Curry 2006). These latter observations support a relatively consistent

enrichment factor between an animal and its diet over the animal’s lifetime.

Previous studies have shown that the nutritional state of the animal can have a

strong influence on ∆15N (Hobson et al. 1993, Doucett et al. 1999, Jardine et al. 2005a).

During starvation or other periods of high nitrogen demand, excretion of concentrated

nitrogenous waste products (e.g. urea) that are isotopically “light” (i.e. low in nitrogen-

15, Steele and Daniel 1978) leads to 15N enrichment in the remaining nitrogen pool

available for anabolism. Animals that are stressed have been shown to have higher δ15N than their non-stressed counterparts, and certain tissues can be more affected than others

(Hobson et al. 1993, Doucett et al. 1999, Jardine et al. 2005a). The mechanism underlying this phenomenon is relatively unknown but is likely related to 15N

fractionation during the transfer of nitrogen (transamination and deamination) between

amino acids in elimination processes (Macko et al. 1986, Bada et al. 1989). The

relationships between nutritional state and δ15N have led to investigations into the effect

of dietary quality on ∆15N. Dietary quality is often defined using elemental ratios of the

diet such as C/N. While theoretically plausible, results to date have been equivocal, with

some authors finding strong associations between diet quality and ∆15N (Hobson and

Clark 1992a, Adams and Sterner 2000), and others finding weak or variable relationships

(Vanderklift and Ponsard 2003, Jardine et al. 2005b, Robbins et al. 2005).

Another complication to the application and use of ∆13C values is the process of

endogenous lipid synthesis, which can affect the ∆13C between an animal and its diet if

the consumer has a higher proportion of lipids when compared to its prey. Lipids are

isotopically lighter due to fractionation against the heavier isotope during the oxidation of

pyruvate to acetyl-CoA during lipid formation (DeNiro and Epstein 1977). Some studies

have used lipid-extracted tissues to remove the effects of this confounding variable (e.g.

Hobson et al. 2002); lipids are removed with a solvent such as chloroform-methanol

(Bligh and Dyer 1959). However, lipid extraction has been shown to influence ∆15N

(Sotiropoulos et al. 2004), possibly due to protein degradation during the extraction

process. In addition, animals that are reared on a constant diet but have different body

compositions or metabolic characteristics can show variations in ∆13C of up to 1.5‰ in

lipid-free matter, suggesting that lipid load is not solely responsible for more negative

δ13C in some animals compared to their diet (Focken and Becker 1998, Gaye-Siessegger

et al. 2004).

An alternative to lipid extraction is lipid normalization using carbon to nitrogen

ratios (C:N) that are often provided by analytical labs alongside stable isotope data.

Percent lipid can be related to C:N in tissues, as for a marine food web (McConnaughey

and McRoy 1979), and the δ13C values can be adjusted accordingly. A similar approach

can be used to normalize contaminant concentrations (Hebert and Keenleyside 1995)

without having to lipid extract tissues for SIA.

The lipid issue is particularly relevant to studies on lipophilic contaminants, and is

confounded by the fact that lipids are highly correlated to both trophic position and

concentrations of these pollutants (e.g. Kucklick and Baker 1998). For this reason, many

studies express concentrations of organic pollutants on a lipid weight basis, thus

accounting for increased % lipid up the food chain, and then use δ15N to describe the

remaining variability in contaminant concentrations among species or individuals (e.g.

Kucklick and Baker 1998, Fisk et al. 2001a).

2.2.3 Consideration/Source of Uncertainty 3: Different Tissues From an

Individual Animal Have Variable Stable Isotope Ratios

If assumptions 1 & 2 are met, i.e. ∆15N and ∆13C have been established and it is

unlikely that the nutritional state of the animal is a concern under the experimental

conditions, the researcher must then choose an appropriate tissue to represent the

animal’s diet. However, because different tissues may have different degrees of

fractionation relative to the diet (Hobson and Clark 1992a) and turnover rates (Hobson

and Clark 1992b), a single tissue may not be adequate in tracing the flow of organic

matter and contaminants through an ecosystem (MacNeil et al. 2005). These two issues

(fractionation and turnover) will be examined separately.

2.2.3.1 Different Fractionation Among Tissues

When animals are on a fixed diet and hence are in isotopic equilibrium, isotopic

ratios may vary amongst tissues due to their differential composition of proteins, lipids

and carbohydrates (Hobson and Clark 1992a). For example, variation (up to 4‰) in

isotope ratios amongst tissues (whole body, muscle, heart and liver) has been reported for

lab-reared rainbow trout (Oncorhynchus mykiss, Pinnegar and Polunin 1999), while

different diet-tissue fractionations of 13C and 15N in blood, liver, muscle, collagen, and feather have been shown in several bird species (Hobson and Clark 1992a). As discussed previously for among-individuals comparisons, the lipid load of a specific tissue may also affect inter-tissue differences in δ13C and will change seasonally depending on diet,

reproductive state and migration patterns. Lipid rich tissues such as liver and eggs tend

to have lower δ13C than low-lipid tissues such as muscle, hair and feathers (Tieszen et al.

1983, Jardine et al. 2005a). Differences in amino acid composition may also determine

the distribution of heavy and light carbon and nitrogen isotopes in different tissues, as

certain amino acids have been shown to be “heavier” than others (Gaebler et al. 1966).

2.2.3.2 Different Turnover Rates Among Tissues

When animals switch diets, the rate of approach of isotopic ratios towards that of

the new diet varies among tissues. Different rates of turnover amongst tissues was first

shown by Tieszen et al. (1983), who reported faster turnover in gerbil liver (half-life =

6.4 days) compared to muscle (half-life = 27.6 days) and hair (half-life = 47.5 days).

These differences were related to differential metabolic activity (or lack thereof) of the

tissues. Different tissues from animals that periodically switch diets in response to

changes in habitat or prey availability may have distinctive isotopic ratios depending on

the turnover rate of the tissue, the duration of the diet switch, and the isotopic

composition of the new diet.

For these reasons, analysis of several tissues for both stable isotopes and

contaminants may offer additional information on the temporal dynamics of contaminant

uptake than would be available from a single tissue (Table 2). Other factors to consider

when choosing tissues include the ability to sample without sacrificing the animal

(Hobson and Clark 1993, Baker et al. 2004) and the mass of the tissue required for SIA

and contaminant analyses. Following the results of Pinnegar and Polunin (1999), most

researchers have settled on muscle tissue as the logical choice for fishes in food web

studies because of its low lipid content, intermediate turnover rate and relevance to fish

consumption by humans; however the liver, which has been shown to exhibit a faster

turnover rate in fishes (McMillan and Houlihan 1989, Perga and Gerdeaux 2005,

MacNeil et al. 2006) may also be important in studies as a center for storage of

contaminants. In birds and mammals, liver turns over more rapidly than other body

tissues (Tieszen et al. 1983, Hobson and Clark 1992b) and therefore could be used to

assess shorter-term dietary changes.

2.2.4 Consideration/Source of Uncertainty 4: Baseline Stable Isotope Ratios Vary

Across Systems

Inputs of nutrients from exogenous sources, both natural and anthropogenic in origin, are common in many aquatic food webs (Polis et al. 1997, Carpenter et al. 1998).

These nutrients often have distinct isotopic ratios (e.g. Chang et al. 2002), their influence

can vary spatially and temporally (Wayland and Hobson 2001), and as a result can cause

primary producers to vary in their isotopic ratios both within and across systems.

Changes in baseline ratios can be due to anthropogenic influences (Cabana and

Rasmussen 1996) such as human sewage and agriculture (Anderson and Cabana 2005), or natural nitrogen transformations, including nitrification and denitrification (Fry 1999).

Locally-different baseline δ13C and especially δ15N have the potential to confound

interpretation of trophic differences within a species when one compares across systems

or over a spatial gradient within a system. For example, despite a relatively low variance

in baseline δ15N among three lakes, sensitivity analysis revealed large potential effects of

these different baselines on TL estimates for higher order consumers (Post 2002). Far

greater effects of varying baselines are expected when comparing organisms from areas

of dense human activity with more pristine locations (Cabana and Rasmussen 1996).

2.2.5 Consideration/Source of Uncertainty 5: The Presence of True Omnivores

While debate continues over the influence of omnivory (defined here as feeding on more than one TL) on the evolutionary stability of consumers in food webs (Pimm and

Lawton 1978, Tanabe and Namba 2005), field studies suggest that there are indeed instances where considerable omnivory does occur and stable isotope data have supported this conclusion (Kling et al. 1992, Cabana and Rasmussen 1994). It is important to consider omnivory in SIA studies because periodic omnivory may lead to a disproportionate uptake of contaminants when compared to any concurrent shifts in isotopic ratios. This may be a consequence of differences in the kinetics of the uptake and excretion of organic matter and contaminants. Bioenergetics models for certain contaminants (e.g. methyl Hg and cesium) predict that fish with high activity and food consumption rates will more rapidly accumulate contaminants than they will accumulate biomass (Rasmussen and Vander Zanden 2004). Since isotopic change following a diet switch in ectotherms is thought to be dominated by growth (Hesslein et al. 1993, Suzuki et al. 2005, but see MacNeil et al. 2006), it follows that a switch in diet by this organism to a prey type with a large contaminant load might generate higher contaminant concentrations for the consumer than predicted based on isotopic ratios alone (Fisk et al.

2002). Likewise for endotherms, half-lives of many contaminants are far greater than the turnover rate of carbon and nitrogen in tissue proteins (Clark et al. 1987, Drouillard and

Norstrom 2003, Dalerum and Angerbjorn 2005), resulting in higher than expected contaminant burdens for, e.g., birds that scavenge marine mammals with large contaminant loads (Fisk et al. 2001b, Hobson et al. 2002). In this type of application,

contaminant information can in fact aid in the interpretation of stable isotope data, as the omnivory may be detected with contaminant data but not with SIA.

2.2.6 Consideration/Source of Uncertainty 6: Movement of Animals and Nutrients

Between Food Webs/Ecosystems

Despite the integrative nature of SIA, collection and analysis of food web components essentially presents a static measure of food web structure and energy and contaminant flow, unless different tissues with different turnover rates are used. This may fail to represent the dynamic nature of temporally variable food webs, particularly those that receive periodic influxes of nutrients, organic matter, and contaminants from adjacent systems (Polis et al. 1997). Certain species may accumulate large body burdens of contaminants that can be deposited in breeding habitats through excretion, predation, and decomposition (Blais et al. 2005). Sarica et al. (2004) demonstrated high methyl Hg

concentrations in invertebrates inhabiting a Lake Ontario, Canada tributary following a

large run of spawning Chinook salmon (Oncorhynchus tshawytscha). Differences in

organochlorine residues among populations of grizzly bears (Ursus arctos horribilis) in

British Columbia, Canada were related to the seasonal consumption of a more

contaminated diet of salmon by one population: this difference in dietary habits was

determined using hair isotopic ratios (Christensen et al. 2005). Coastal dwelling bears

that shifted diets to salmon had higher concentrations of ∑DDT, ∑CHL, and dieldrin than

their interior counterparts that were almost exclusively herbivorous (Christensen et al.

2005).

A second challenge in this approach arises from the presence of migratory species

that arrive from habitats with higher or lower baseline contaminant concentrations or

isotope ratios; these individuals therefore appear as outliers in contaminant vs. δ15N plots or in dual isotope plots. For example, higher-than-expected concentrations of persistent organic pollutants in black-legged kittiwakes (Rissa tridactyla) in the Canadian Arctic were presumed to stem from the accumulation of these pollutants from their wintering grounds on the eastern North American seaboard (Fisk et al. 2001a). Similarly, residents and (tributary) migrants in a population of American dippers (Cinclus mexicanus), identified using SIA, had significantly different contaminant concentrations (Morrissey et al. 2004). Resident dippers ate significantly more fish than the migrants, leading to higher Hg, organochlorines, and PCBs in the former cohort (Morrissey et al. 2004).

Perhaps the best examples of complex systems with constant species movements and multiple food sources are estuaries (Deegan and Garritt 1997). Species found in the estuary at any particular time may spend significant time outside the estuary in freshwater or marine water where contaminants and stable isotope ratios likely differ. Interestingly, there have been few published studies that have examined food web influences on contaminants in estuaries, which may be due in large part to the complexity of these systems, with seasonal use by several species as migratory pathways or nursery areas.

For example, few significant relationships between PCB congener concentrations and TL were found for a Georgia, USA, estuarine food web, despite high PCB concentrations in the biota (Smith 2005). Likewise, only weak relationships between Hg and TL were found for a New Brunswick, Canada, estuarine food web (Pastershank 2001).

Understanding nutrient and organic matter exchange within estuaries may therefore be

important in assessing contaminant transfer on a larger scale (Polis et al. 1997).

2.3 New Directions

Over the past few years, SIA has been used in new ways to study bioaccumulation

and biomagnification. Because contaminant concentrations in biota are affected by the

complex interplay between several environmental processes, researchers have begun to

use isotopic ratios in conjunction with physical, chemical or biological factors to generate

more sophisticated models to understand contaminant concentrations in consumers

(Greenfield et al. 2001, Evenset et al. 2004, Evans et al. 2005). A comprehensive analysis of biomagnification slopes from δ15N vs. contaminant associations available in

the literature (Broman et al. 1992, Kidd 1998) may prove useful in understanding factors, such as contaminant properties, food web structure, and productivity, that lead to higher than normal rates of biomagnification in aquatic systems.

The effect of the arrival of invasive species on contaminant concentrations can also be assessed using SIA (Cabana and Rasmussen 1994, Vander Zanden and

Rasmussen 1996). One such species, rainbow smelt, has been shown to lengthen food chains (Vander Zanden and Rasmussen 1996, Swanson et al. 2003). However, while

Vander Zanden and Rasmussen (1996) reported that smelt invasion resulted in higher Hg concentrations in lake trout, Swanson et al. (2003) found similar Hg concentrations in the top predator walleye (Sander vitreus) in smelt-invaded and reference lakes, suggesting that walleye were not preferentially feeding on smelt or other factors, such as ecosystem

productivity, were interacting with trophic transfer to control Hg biomagnification (Kidd

et al. 1999).

Assessing toxicological effects and dietary habits concurrently will improve the

ability to determine the ecological relevance of changes in food pathways for those

contaminants for which dietary exposure is significant. For example, marine consumers

connected to a clam-based food web accumulated selenium more rapidly through the

food web (higher selenium-δ15N slopes) than those animals in a food web supported by crustaceans within the same region (Stewart et al. 2004). This steep biomagnification slope in the clam-based food web pushed several species above the toxicity threshold for selenium (Stewart et al. 2004). Other links between isotope ratios and toxic effects have been observed in the lab. Snowy egrets (Egreta thulla) exposed to high levels of methyl

Hg had tissue-specific (muscle and liver) shifts in isotopic ratios that were likely associated with protein stress and degradation (Shaw-Allen et al. 2005), demonstrating that not only can isotopes indicate likely exposure to contaminants, contaminants can also affect isotope ratios.

Analysis of archived tissues could provide information on spatial and temporal trends in contaminant concentrations (e.g. museum collections, Rocque and Winker

2005). These time series will allow us to determine if environmental concentrations are increasing or decreasing, particularly in relation to bureaucratic decisions that regulated releases, such as the banning of DDT in North America (Braune et al. 2001) or the introduction of the Clean Air Act in the United States. Pairing the archived contaminant data with stable isotope data of the same tissues (Thompson et al. 1998), provided baseline δ15N has not changed, may allow researchers to account for long-term changes

in trophic structuring that could alter contaminant profiles in consumers (Hobson et al.

1997).

The use of compound-specific stable isotope analyses, both in the tissue fraction

of consumers (Herman et al. 2005) and in the contaminants themselves (Horii et al.

2005), is gaining popularity and will no doubt add to the understanding of contaminant

cycling within organisms and through food webs, and to the ability to link contaminants in consumers to their original source.

SIA has had an almost exponential increase in its use in ecology and ecotoxicological studies over the past several decades (Kelly 2000), and its number of applications will no doubt continue to expand. In order to improve interpretation of stable isotope data in ecotoxicology studies, the following is recommended:

1. When using SIA across multiple systems it is critical to standardize the δ15N of an organism to the basal value within each system to avoid inappropriate conclusions with respect to relative TL. Typical variation in baseline δ15N amongst ecosystems

(range of ~12‰, Cabana and Rasmussen 1996) outlined in consideration 4 exceeds the variability associated with considerations 1-3: (1) interspecific differences in ∆15N

(range of ~5‰ Post 2002), (2) intraspecific differences in ∆15N associated with the

physiology of the animal (range of ~5‰, Hobson et al. 1993, Adams and Sterner 2000),

or (3) variability among tissues within an animal (range of ~5‰, DeNiro and Epstein

1981, Pinnegar and Polunin 1999, Vanderklift and Ponsard 2003). Indeed considerations

2 and 3 are embedded within consideration 1, in that much of the variability amongst

species is likely related to differences in physiology and tissues sampled (Vanderklift and

Ponsard 2003).

Proper procedures for assessing baselines in lakes have been outlined (Post 2002),

and it is recommended to use SIA of filter feeding mussels to quantify the pelagic

baseline and snails that feed on benthic algae to establish the littoral baseline. This was

an advance from the use of zooplankton that can have considerable interspecific or

seasonal variation in isotopic ratios within lakes (Matthews and Mazumder 2003).

Unfortunately, proper techniques for baseline assessment in small fluvial systems are less

well developed, in part due to the rarity of long-lived primary consumers in these habitats

(but see Howard et al. 2005). Further development of baseline assessment, using a variety of functional feeding groups (e.g. scrapers and shredders, Cummins and Klug

1979) shows promise for future stable isotope and ecotoxicology studies in streams and rivers (Anderson and Cabana 2007).

2. When possible, in order to assess species-specific diet-tissue fractionation, study animals should be reared on a constant diet in a laboratory setting. If this is not feasible, use available information from the literature and choose the most appropriate values for ∆15N and ∆13C (Post 2002, Vander Zanden and Rasmussen 2001, McCutchan et al. 2003, Vanderklift and Ponsard 2003, Minagawa and Wada 1984).

3. If available, incorporate other biological information into the study, particularly

animal movement patterns and short-term shifts in diet (e.g. omnivory). Supplementation

of isotope data with more traditional measures of feeding behaviour, such as gut content

analysis (GCA) and visual observations, as well as analysis of multiple tissues for SIA,

will enhance understanding of contaminant data that may not have been revealed with

isotope analysis of a single tissue (Fisk et al. 2002). For the most part, past studies have

found relatively good agreement between GCA and SIA (Davenport and Bax 2002, Grey et al. 2002)

By considering the sources of uncertainty outlined in this paper and by continuing to conduct controlled laboratory studies that better define isotope kinetics, future research with SIA will continue to provide new insights into factors that govern the movement of organic matter and contaminants through ecosystems.

2.4 Acknowledgements

Funding for this review was provided by the Canada Research Chairs Program, the

Natural Sciences and Engineering Research Council, and the New Brunswick

Environmental Trust Fund.

2.5 References

Adams, T.S., and Sterner, R.W. 2000. The effect of dietary nitrogen content on trophic

level 15N enrichment. Limnology and Oceanography 45: 601-607.

Anderson, C., and Cabana, G. 2005. δ15N in riverine food webs: effects of N inputs from

agricultural watersheds. Canadian Journal of Fisheries and Aquatic Sciences 62:

333-340.

Anderson, C., and Cabana, G. 2007. Estimating the trophic position of aquatic consumers

in river food webs using stable nitrogen isotopes. Journal of the North American

Benthological Society 26: 273-285.

Atwell, L., Hobson, K.A., and Welch, H.E. 1998. Biomagnification and bioaccumulation

of mercury in an arctic marine food web: insights from stable nitrogen isotope

analysis. Canadian Journal of Fisheries and Aquatic Sciences 55: 1114-1121.

Bada, J.L., Schoeninger, M.J., and Schimmelmann, A. 1989. Isotopic fractionation during

peptide bond hydrolysis. Geochimica et Cosmochimica Acta 53: 3337-3341.

Baker, R.F., Blanchfield, P.J., Paterson, M.J., Flett, R.J., and Wesson, L. 2004.

Evaluation of nonlethal methods for the analysis of mercury in fish tissue.

Transactions of the American Fisheries Society 133: 568-576.

Bearhop, S., Waldron, S., Thompson, D., and Furness, R. 2000. Bioamplification of

mercury in great skua Catharacta skua chicks: the influence of trophic status as

determined by stable isotope signatures of blood and feathers. Marine Pollution

Bulletin 40: 181-185.

Ben-David, M., Flynn, R.W., and Schell, D.M. 1997. Annual and seasonal changes in the

diet of martens: evidence from stable isotope analysis. Oecologia 111: 280-291.

Berglund, O., Larsson, P., and Broman, D. 2001. Organochlorine accumulation and stable

isotope ratios in an Atlantic salmon (Salmo salar) population from the Baltic Sea.

The Science of the Total Environment 281: 141-151.

Berglund, O., Nystrom, P., and Larsson, P. 2005. Persistent organic pollutants in river

food webs: influence of trophic position and degree of heterotrophy. Canadian

Journal of Fisheries and Aquatic Sciences 62: 2021-2032.

Blais, J.M., Kimpe, L.E., McMahon, D., Keatley, B.E., Mallory, M.L., Douglas, M.S.V.,

and Smol, J.P. 2005. Arctic seabirds transport marine-derived contaminants.

Science 309: 445.

Bligh, E.G., and Dyer, W.J. 1959. A rapid method of total lipid extraction and

purification. Canadian Journal of Biochemistry and Physiology 37: 911-917.

Borga, K., Gabrielsen, G.W., and Skaare, J.U. 2002a. Differences in contamination load

between pelagic and sympagic invertebrates in the arctic marginal ice zone:

influence of habitat, diet and geography. Marine Ecology Progress Series 325: 157-

169.

Borga, K., Gulliksen, B., Gabrielsen, G.W., and Skaare, J.U. 2002b. Size-related

bioaccumulation and between-year variation of organochlorines in ice-associated

amphipods from the Arctic Ocean. Chemosphere 46: 1383-1392.

Borga, K., Fisk, A.T., Hoekstra, P.F., and Muir, D.C.G. 2004. Biological and chemical

factors of importance in the bioaccumulation and trophic transfer of persistent

organochlorine contaminants in arctic marine food webs. Environmental

Toxicology and Chemistry 23: 2367-2385.

Braune, B.M., Donaldson, G.M., and Hobson, K.A. 2001. Contaminant residues in

seabird eggs from the Canadian Arctic. I. Temporal trends 1975-1998.

Environmental Pollution 114: 39-54.

Braune, B.M., Donaldson, G.M., and Hobson, K.A. 2002. Contaminant residues in

seabird eggs from the Canadian Arctic. II. Spatial trends and evidence from stable

isotopes for intercolony differences. Environmental Pollution 117: 133-145.

Broman, D., Naf, C., Rolff, C., Zebuhr, Y., Fry, B., and Hobbie, J. 1992. Using ratios of

stable nitrogen isotopes to estimate bioaccumulation and flux of polychlorinated

dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) in two food chains from

the northern Baltic. Environmental Toxicology and Chemistry 11: 331-345.

Cabana, G., and Rasmussen, J.B. 1994. Modelling food chain structure and contaminant

bioaccumulation using stable nitrogen isotopes. Nature 372: 255-257.

Cabana, G., Tremblay, A., Kalff, J., and Rasmussen, J.B. 1994. Pelagic food chain

structure in Ontario lakes: a determinant of mercury levels in lake trout (Salvelinus

namaycush). Canadian Journal of Fisheries and Aquatic Sciences 51: 381-389.

Cabana, G., and Rasmussen, J.B. 1996. Comparison of aquatic food chains using nitrogen

isotopes. Proceedings of the National Academy of Sciences, U.S.A. 93: 10844-

10847.

Campbell, L.M., Schindler, D.W., Muir, D.C.G., Donald, D.B., and Kidd, K.A. 2000.

Organochlorine transfer in the food web of subalpine Bow Lake, Banff National

Park. Canadian Journal of Fisheries and Aquatic Sciences 57: 1258-1269.

Campbell, L.M., Fisk, A.T., Wang, X., Kock, G., and Muir, D.C.G. 2005a. Evidence for

biomagnification of rubidium in freshwater and marine food webs. Canadian

Journal of Fisheries and Aquatic Sciences 62: 1161-1167.

Campbell, L.M., Norstrom, R.J., Hobson, K.A., Muir, D.C.G., Backus, S., and Fisk, A.T.

2005b. Mercury and other trace elements in a pelagic Arctic marine food web

(Northwater Polynya, Baffin Bay). The Science of the Total Environment 351-352:

247-263.

Carmusso, M., Martinotti, W., Balestrini, R., and Guzzi, L. 1998. C and N stable isotopes

and trace metals in selected organisms from the River Po Delta. Chemosphere 37:

2911-2920.

Carpenter, S.R., Caraco, N.F., Correll, D.L., Howarth, R.W., Sharpley, A.N., and Smith,

V. 1998. Nonpoint pollution of surface waters with phosphorus and nitrogen.

Ecological Applications 8: 559-568.

Chang, C.C.Y., Kendall, C., Silva, S.R., Battaglin, W.A., and Campbell, D.H. 2002.

Nitrate stable isotopes: tools for determining nitrate sources among different land

uses in the Mississippi River Basin. Canadian Journal of Fisheries and Aquatic

Sciences 59: 1874-1885.

Christensen, J.R., MacDuffie, M., MacDonald, R.W., Whiticar, M., and Ross, P.S. 2005.

Persistent organic pollutants in British Columbia grizzly bears: consequence of

divergent diets. Environmental Science and Technology 39: 6952-6960.

Clark, T.P., Norstrom, R.J., Fox, G.A., Won, H.T. 1987. Dynamics of organochlorine

compounds in herring gulls (Larus argentatus): II. A two-compartment model and

data for ten compounds. Environmental Toxicology and Chemistry 6: 547-559.

Croteau, M.-N., Luoma, S.N., and Stewart, A.R. 2005. Trophic transfer of metals along

freshwater food webs: Evidence of cadmium biomagnification in nature. Limnology

and Oceanography 50: 1511-1519.

Cummins, K.W., and Klug, M.J. 1979. Feeding ecology of stream invertebrates. Annual

Review of Ecology and Systematics 10: 147-172.

Dalerum, F., and Angerbjorn, A. 2005. Resolving temporal variation in vertebrate diets

using naturally occurring stable isotopes. Oecologia 144: 647-658.

Das, K., Lepoint, G., Loizeau, V., Debacker, V., Dauby, P., and Bouquegneau, J.M.

2000. Tuna and dolphin associations in the North-east Atlantic: Evidence of

different ecological niches from stable isotope and heavy metal measurements.

Marine Pollution Bulletin 40: 102-109.

Davenport, S.R., and Bax, N.J. 2002. A trophic study of a marine ecosystem off

southeastern Australia using stable isotopes of carbon and nitrogen. Canadian

Journal of Fisheries and Aquatic Sciences 59: 514-530.

Deegan, L.A., and Garritt, R.H. 1997. Evidence for spatial variability in estuarine food

webs. Marine Ecology Progress Series 147: 31-47.

DeNiro, M.J., and Epstein, S. 1977. Mechanism of carbon isotope fractionation

associated with lipid synthesis. Science 197: 261-263.

DeNiro, M.J., and Epstein, S. 1978. Influence of the diet on the distribution of carbon

isotopes in animals. Geochimica et Cosmochimica Acta 42: 495-506.

DeNiro, M.J., and Epstein, S. 1981. Influence of the diet on the distribution of nitrogen

isotopes in animals. Geochimica et Cosmochimica Acta 45: 341-351.

Doucett, R.R., Booth, R.K., Power, G., and McKinley, R.S. 1999. Effects of the

spawning migration on the nutritional status of anadromous Atlantic salmon (Salmo

salar): insights from stable-isotope analysis. Canadian Journal of Fisheries and

Aquatic Sciences 56: 2172-2180.

Drouillard, K.G., and Norstrom, R.J. 2003. The influence of diet properties and feeding

rates on PCB toxicokinetics in the ring dove. Archives of Environmental

Contamination and Toxicology 44: 97-106.

Dunton, K.H., and Schell, D.M. 1987. Dependence of consumers on macroalgal

(Laminaria solidungula) carbon in an Arctic kelp community: δ13C evidence.

Marine Biology 93: 615-625.

Evans, M.S., Lockhart, W.L., Doetzel, L., Low, G., Muir, D., Kidd, K., Stephens, G., and

Delaronde, J. 2005. Elevated mercury concentrations in fish in lakes in the

MacKenzie River Basin: The role of physical, chemical, and biological factors. The

Science of the Total Environment 351-352: 479-500.

Evenset, A., Christensen, G.N., Skotvold, T., Fjeld, E., Schlabach, M., Wartena, E., and

Gregor, D. 2004. A comparison of organic contaminants in two high Arctic lake

ecosystems, Bjornoya (Bear Island), Norway. The Science of the Total

Environment 318: 125-141.

Fisk, A.T., Hobson, K.A., and Norstrom, R.J. 2001a. Influence of chemical and

biological factors on trophic transfer of persistent organic pollutants in the

Northwater Polynya marine food web. Environmental Science and Technology 35:

732-738.

Fisk, A.T., Moisey, J., Hobson, K.A., Karnovsky, N.J., and Norstrom, R.J. 2001b.

Chlordane components and metabolites in seven species of Arctic seabirds from the

Northwater Polynya: relationships with stable isotopes of nitrogen and enantiomeric

fractions of chiral components. Environmental Pollution 113: 225-238.

Fisk, A.T., Tittlemier, S.A., Pranschke, J.L., and Norstrom, R.J. 2002. Using

anthropogenic contaminants and stable isotopes to assess the feeding ecology of

Greenland sharks. Ecology 83: 2162-2172.

Fisk, A.T., Hoekstra, P.F., Gagnon, J.-M., Duffe, J., Norstrom, R.J., Hobson, K.A.,

Kwan, M., and Muir, D.C.G. 2003. Influence of habitat, trophic ecology and lipids

on, and spatial trends of, organochlorine contaminants in Arctic marine

invertebrates. Marine Ecology Progress Series 262: 201-214.

Focken, U., and Becker, K. 1998. Metabolic fractionation of stable carbon isotopes:

implications of different proximate compositions for studies of the aquatic food

webs using δ13C data. Oecologia 115: 337-343.

Fox, G.A., Grasman, K.A., Hobson, K.A., Williams, K., Jeffrey, D., and Hanbridge, B.

2002. Contaminant residues in tissues of adult and prefledged herring gulls from the

Great Lakes in relation to diet in the early 1990s. Journal of Great Lakes Research

28: 643-663.

France, R.L. 1995. Carbon-13 enrichment in benthic compared to planktonic algae:

foodweb implications. Marine Ecology Progress Series 124: 307-312.

Fry, B. 1999. Using stable isotopes to monitor watershed influences on aquatic

trophodynamics. Canadian Journal of Fisheries and Aquatic Sciences 56: 2167-

2171.

Fry, B., and Sherr, E.B. 1984. δ13C measurements as indicators of carbon flow in marine

and freshwater ecosystems. Contributions in Marine Science 27: 13-47.

Gaebler, O.H., Vitti, T.G., and Vukmirovich, R. 1966. Isotope effects in metabolism of

14N and 15N from unlabeled dietary proteins. Canadian Journal of Biochemistry

44: 1249-1257.

Gaines, K.F., Romanek, C.S., Boring, C.S., Lord, C.G., Gochfeld, M., and Burger, J.

2002. Using raccoons as an indicator species for metal accumulation across trophic

levels: a stable isotope approach. Journal of Wildlife Management 66: 811-821.

Gaye-Siessegger, J., Focken, U., Muetzel, S., Abel, H., and Becker, K. 2004. Feeding

level and individual metabolic rate affect δ13C and δ15N values in carp:

implications for food web studies. Oecologia 138: 175-183.

Greenfield, B.K., Hrabik, T.R., Harvey, C.J., and Carpenter, S.R. 2001. Predicting

mercury levels in yellow perch: use of water chemistry, trophic ecology, and spatial

traits. Canadian Journal of Fisheries and Aquatic Sciences 58: 1419-1429.

Grey, J., Thackeray, S.J., Jones, R.I., and Shine, A. 2002. Ferox trout (Salmo trutta) as

'Russian dolls': complementary gut content and stable isotope analyses of the Loch

Ness foodweb. Freshwater Biology 47: 1235-1243.

Gustin, M.S., Saito, L., and Peacock, M. 2005. Anthropogenic impacts on mercury

concentrations and nitrogen and carbon isotope ratios in fish muscle tissue of the

Truckee River watershed, Nevada, USA. Science of the Total Environment 347:

282-294.

Hall, B.D., R.A. Bodaly, R.J.P. Fudge, J.W.M. Rudd, and D.M. Rosenberg 1997. Food as

the dominant pathway of methylmercury uptake by fish. Water, Air, and Soil

Pollution 100: 13-24.

Hebert, C.E., and K.A. Keenleyside 1995. To normalize or not to normalize? Fat is the

question. Environmental Toxicology and Chemistry 14: 801-807.

Hecky, R.E., and Hesslein, R.H. 1995. Contributions of benthic algae to lake food webs

as revealed by stable isotope analysis. Journal of the North American Benthological

Society 14: 631-653.

Herman, D.P., Burrows, D.G., Wade, P.R., Durban, J.W., Matkin, C.O., LeDuc, R.G.,

Barrett-Lennard, L.G., and Krahn, M.M. 2005. Feeding ecology of eastern North

Pacific killer whales Orcinus orca from fatty acid, stable isotope, and

organochlorine analyses of blubber biopsies. Marine Ecology Progress Series 302:

275-291.

Hesslein, R.H., Hallard, K.H., and Ramlal, P. 1993. Replacement of sulfur, carbon, and

nitrogen in tissue of growing broad whitefish (Coregonus nasus) in response to a

change in diet traced by δ34S, δ13C, and δ15N. Canadian Journal of Fisheries and

Aquatic Sciences 50: 2071-2076.

Hobson, K.A. 1999. Tracing origins and migration of wildlife using stable isotopes: a

review. Oecologia 120: 314-326.

Hobson, K.A., and Clark, R.G. 1992a. Assessing avian diets using stable isotopes II:

factors influencing diet-tissue fractionation. The Condor 94: 189-197.

Hobson, K.A., and Clark, R.G. 1992b. Assessing avian diets using stable isotopes I:

turnover of 13C in tissues. The Condor 94: 181-188.

Hobson, K.A., and Clark, R.G. 1993. Turnover of 13C in cellular and plasma fractions of

blood: implications for nondestructive sampling in avian dietary studies. The Auk

110: 638-641.

Hobson, K.A., and Welch, H.E. 1992. Determination of trophic relationships within a

high Arctic marine food web using δ13C and δ15N analysis. Marine Ecology

Progress Series 84: 9-18.

Hobson, K.A., Alisauskas, R.T., and Clark, R.G. 1993. Stable-nitrogen isotope

enrichment in avian tissues due to fasting and nutritional stress: implications for

isotopic analyses of diet. The Condor 95: 388-394.

Hobson, K.A., Hughes, K.D., and Ewins, P.J. 1997. Using stable-isotope analyses to

identify endogenous and exogenous sources of nutrients in eggs of migratory birds:

Applications to Great Lakes contaminants research. The Auk 114: 467-478.

Hobson, K.A., Fisk, A., Karnovsky, N., Holst, M., Gagnon, J.-M., and Fortier, M. 2002.

A stable isotope (δ13C, δ15N) model for the North Water food web: implications

for evaluating trophodynamics and the flow of energy and contaminants. Deep-Sea

Research II 49: 5131-5150.

Hoekstra, P.F., Braune, B.M., O'Hara, T.M., Elkin, B., Solomon, K.R., and Muir, D.C.G.

2003. Organochlorine contaminant and stable isotope profiles in Arctic fox (Alopex

lagopus) from the Alaskan and Canadian Arctic. Environmental Pollution 122: 423-

433.

Horii, Y., Kannan, K., Petrick, G., Gamo, T., Falandysz, J., and Yamashita, N. 2005.

Congener-specific carbon isotopic analysis of technical PCB and PCN mixtures

using two-dimensional gas chromatography – Isotope ratio mass spectrometry.

Environmental Science and Technology 39: 4206-4212.

Howard, J.K., Cuffey, K.M., and Solomon, M. 2005. Toward using Margaritifera falcata

as an indicator of base level nitrogen and carbon isotope ratios: insights from two

California Coast Range rivers. Hydrobiologia 541: 229-236.

Jardine, T.D., and Cunjak, R.A. 2005. Analytical error in stable isotope ecology.

Oecologia 144: 528-533.

Jardine, T.D., and Curry, R.A. 2006. Unique perspectives on the influence of size and age

on consumer δ15N from a rainbow smelt complex. Journal of Fish Biology 69: 215-

223.

Jardine, T.D., Gray, M.A., McWilliam, S.M., and Cunjak, R.A. 2005a. Stable isotope

variability in tissues of temperate stream fishes. Transactions of the American

Fisheries Society 134: 1103-1110.

Jardine, T.D., Curry, R.A., Heard, K.S., and Cunjak, R.A. 2005b. High fidelity: Isotopic

relationship between stream invertebrates and their stomach contents. Journal of the

North American Benthological Society 24: 290-299.

Jarman, W.M., Hobson, K.A., Sydeman, W.J., Bacon, C.E., and McLaren, E.B. 1996.

Influence of trophic position and feeding location on contaminant levels in the Gulf

of the Farallones food web revealed by stable isotope analysis. Environmental

Science and Technology 30: 654-660.

Kelly, J.F. 2000. Stable isotopes of carbon and nitrogen in the study of avian and

mammalian trophic ecology. Canadian Journal of Zoology 78: 1-27.

Kidd, K.A. 1998. Use of stable isotope ratios in freshwater and marine biomagnification

studies. In Environmental Toxicology: Current Developments. Edited by J. Rose.

Gordon & Breach Science Publishers, Amsterdam pp. 357-376.

Kidd, K.A., Schindler, D.W., Muir, D.C.G., Lockhart, W.L., and Hesslein, R.H. 1995a.

High concentrations of toxaphene in fishes from a subarctic lake. Science 269: 240-

242.

Kidd, K.A., Schindler, D.W., Hesslein, R.H., and Muir, D.C.G. 1995b. Correlation

between stable nitrogen isotope ratios and concentrations of organochlorines in

biota from a freshwater food web. The Science of the Total Environment 160/161:

381-390.

Kidd, K.A., Hesslein, R.H., Ross, B.J., Koczanski, K., Stephens, G.R., and Muir, D.C.G.

1998a. Bioaccumulation of organochlorines through a remote freshwater food web

in the Canadian Arctic. Environmental Pollution 102: 91-103.

Kidd, K.A., Schindler, D.W., Hesslein, R.H., and Muir, D.C.G. 1998b. Effects of trophic

position and lipid on organochlorine concentrations in fishes from subarctic lakes in

Yukon Territory. Canadian Journal of Fisheries and Aquatic Sciences 55: 869-881.

Kidd, K.A., Paterson, M.J., Hesslein, R.H., Muir, D.C.G., and Hecky, R.E. 1999. Effects

of northern pike (Esox lucius) additions on pollutant accumulation and food web

structure, as determined by δ13C and δ15N, in a eutrophic and an oligotrophic lake.

Canadian Journal of Fisheries and Aquatic Sciences 56: 2193-2202.

Kidd, K.A., Bootsma, H.A., Hesslein, R.H., Muir, D.C.G., and Hecky, R.E. 2001.

Biomagnification of DDT through the benthic and pelagic food webs of Lake

Malawi, East Africa: Importance of trophic level and carbon source. Environmental

Science and Technology 35: 14-20.

Kidd, K.A., Bootsma, H.A., Hesslein, R.H., Lockhart, W.L., and Hecky, R.E. 2003.

Mercury concentrations in the food web of Lake Malawi, East Africa. Journal of

Great Lakes Research 29 (Suppl. 2): 258-266.

Kiriluk, R.M., Servos, M.R., Whittle, D.M., Cabana, G., and Rasmussen, J.B. 1995.

Using ratios of stable nitrogen and carbon isotopes to characterize the

biomagnification of DDE, mirex, and PCB in a Lake Ontario pelagic food web.

Canadian Journal of Fisheries and Aquatic Sciences 52: 2660-2674.

Kling, G.W., Fry, B., and O'Brien, W.J. 1992. Stable isotopes and planktonic trophic

structure in Arctic lakes. Ecology 73: 561-566.

Kucklick, J.R. and Baker, J.E. 1998. Organochlorines in Lake Superior's food web.

Environmental Science and Technology 32: 1192-1198.

Mackay, D., and Fraser, A. 2000. Bioaccumulation of persistent organic chemicals:

mechanisms and models. Environmental Pollution 110: 375-391.

Mackintosh, C.E., Maldonado, J., Hongwu, J., Hoover, N., Chong, A., Ikonomou, M.G.,

and Gobas, F.A.P.C. 2004. Distribution of phthalate esters in a marine aquatic food

web: comparison to polychlorinated biphenyls. Environmental Science and

Technology 38: 2011-2020.

Macko, S.A., Fogel Estep, M.L., Engel, M.H., and Hare, P.E. 1986. Kinetic fractionation

of stable nitrogen isotopes during amino acid transamination. Geochimica et

Cosmochimica Acta 50: 2143-2146.

MacNeil, M.A., Skomal, G.B., and Fisk, A.T. 2005. Stable isotopes from multiple tissues

reveal diet switching in sharks. Marine Ecology Progress Series 302: 199-206.

MacNeil, M.A., Drouillard, K.G., and Fisk, A.T. 2006. Variable uptake and elimination

of stable nitrogen isotopes between tissues in fish. Canadian Journal of Fisheries

and Aquatic Sciences 63: 345-353.

Matthews, B., and Mazumder, A. 2003. Compositional and interlake variability of

zooplankton affect baseline stable isotope signatures. Limnology and Oceanography

48: 1977-1987.

Mazak, E.J., MacIsaac, H.J., Servos, M.R., and Hesslein, R. 1997. Influence of feeding

habits on organochlorine contaminant accumulation in waterfowl on the Great

Lakes. Ecological Applications 7: 1133-1143.

McConnaughey, T., and McRoy, C.P. 1979. Food-web structure and the fractionation of

carbon isotopes in the Bering Sea. Marine Biology 53: 257-262.

McCutchan, J.H.Jr., Lewis, W.M.Jr., Kendall, C., and McGrath, C.C. 2003. Variation in

trophic shift for stable isotope ratios of carbon, nitrogen, and sulfur. Oikos 102:

378-390.

McMillan, D.N., and Houlihan, D.F. 1989. Short-term responses of protein synthesis to

re-feeding in rainbow trout. Aquaculture 79: 37-46.

Minagawa, M., and Wada, E. 1984. Stepwise enrichment of 15N along food chains:

further evidence and the relation between δ15N and animal age. Geochimica et

Cosmochimica Acta 48: 1135-1140.

Moisey, J., Fisk, A.T., Hobson, K.A., and Norstrom, R.J. 2001. Hexachlorocyclohexane

(HCH) isomers and chiral signatures of α-HCH in the Arctic marine food web of

the Northwater Polynya. Environmental Science and Technology 35: 1920-1927.

Morrissey, C.A., Bendell-Young, L.I., and Elliott, J.E. 2004. Linking contaminant

profiles to the diet and breeding location of American dippers using stable isotopes.

Journal of Applied Ecology 41: 502-512.

Muir, D.C.G., Segstro, M.D., Hobson, K.A., Ford, C.A., Stewart, R.E.A., and Olpinski,

S. 1995. Can seal eating explain elevated levels of PCBs and organochlorine

pesticides in walrus blubber from eastern Hudson Bay (Canada)? Environmental

Pollution 90: 335-348.

Nisbet, I.C.T., Montoya, J.P., Burger, J., and Hatch, J.J. 2002. Use of stable isotopes to

investigate individual differences in diets and mercury exposures among common

terns Sterna hirundo in breeding and wintering grounds. Marine Ecology Progress

Series 242: 267-274.

O’Leary, M.H. 1988. Carbon isotopes in photosynthesis. BioScience 38: 328-336.

Overman, N., and Parrish, D.L. 2001. Stable isotope composition of walleye: 15N

accumulation with age and area-specific differences in δ13C. Canadian Journal of

Fisheries and Aquatic Sciences 58: 1253-1260.

Pastershank, G.M. 2001. Unifying ecosystem concepts and mercury biomagnification in

an estuarine environment using stable isotopes (δ13C and δ15N). PhD thesis,

University of Ottawa.

Perga, M.E., and Gerdeaux, D. 2005. 'Are fish what they eat' all year round? Oecologia

144: 598-606.

Phillips, D.L., and Gregg, J.W. 2001. Uncertainty in source partitioning using stable

isotopes. Oecologia 127: 171-179.

Phillips, D.L., and Koch, P.L. 2002. Incorporating concentration dependence in stable

isotope mixing models. Oecologia 130: 114-125.

Phillips, D.L., and Gregg, J.W. 2003. Source partitioning using stable isotopes: coping

with too many sources. Oecologia 136: 261-269.

Pimm, S.L., and Lawton, J.H. 1978. On feeding on more than one trophic level. Nature

275: 542-544.

Pinnegar, J.K., and Polunin, N.V.C. 1999. Differential fractionation of δ13C and δ15N

among fish tissues: implications for the study of trophic interactions. Functional

Ecology 13: 225-231.

Polis, G.A., Anderson, W.B., and Holt, R.D. 1997. Toward an understanding of

landscape and food web ecology: The dynamics of spatially subsidized food webs.

Annual Review of Ecology and Systematics 28: 289-316.

Post, D.M. 2002. Using stable isotopes to estimate trophic position: models, methods, and

assumptions. Ecology 83: 703-718.

Power, M., Klein, G.M., Guiguer, K.R.R.A., and Kwan, M.K.H. 2002. Mercury

accumulation in the fish community of a sub-Arctic lake in relation to trophic

position and carbon sources. Journal of Applied Ecology 39: 819-830.

Power, M., Guiguer, K.R.R.A., and Barton, D.R. 2003. Effects of temperature on isotopic

enrichment in Daphnia magna: implications for aquatic food-web studies. Rapid

Communications in Mass Spectrometry 17: 1619-1625.

Quinn, M.R., Feng, X., Folt, C.L., and Chamberlain, C.P. 2003. Analyzing trophic

transfer of metals in stream food webs using nitrogen isotopes. The Science of the

Total Environment 317: 73-89.

Rasmussen, J.B., and Vander Zanden, M.J. 2004. The variation of lake food webs across

the landscape and its effect on contaminant dynamics. In Food webs at the

landscape level. Edited by G.A. Polis, M.E. Power, and G.R. Huxel. The University

of Chicago Press, Chicago pp. 169-182.

Rasmussen, J.B., Rowan, D.J., Lean, D.R.S., and Carey, J.H. 1990. Food chain structure

in Ontario lakes determines PCB levels in lake trout (Salvelinus namaycush) and

other pelagic fish. Canadian Journal of Fisheries and Aquatic Sciences 47: 2030-

2038.

Robbins, C.T., Felicetti, L.A., and Sponheimer, M. 2005. The effect of dietary protein

quality on nitrogen isotope discrimination in mammals and birds. Oecologia 144:

534-540.

Rocque, D.A., and Winker, K. 2005. Use of bird collections in contaminant and stable-

isotope studies. The Auk 122: 990-994.

Rounick, J.S., and Winterbourn, M.J. 1986. Stable carbon isotopes and carbon flow in

ecosystems. BioScience 36: 171-177.

Russell, R.W., Lazar, R., and Haeffner, G.D. 1995. Biomagnification of organochlorines

in Lake Erie white bass. Environmental Toxicology and Chemistry 14: 719-724.

Sagerup, K., Henriksen, E.O., Skaare, J.U., and Gabrielsen, G.W. 2002. Intraspecific

variation in trophic feeding levels and organochlorine concentrations in glaucous

gulls (Larus hyperboreus) from Bjornoya, the Barents Sea. Ecotoxicology 11: 119-

125.

Sarica, J., Amyot, M., Hare, L., Doyon, M.-R., and Stanfield, L. 2004. Salmon derived

mercury and nutrients in a Lake Ontario spawning stream. Limnology and

Oceanography 49: 891-899.

Smith, J.S. 2005. Trophic transfer of organochlorines in a salt marsh estuary. MSc thesis,

University of Georgia.

Shaw-Allen, P.L., Romanek, C.S., Bryan, A.L.Jr., Brant, H., and Jagoe, C.H. 2005. Shifts

in relative tissue δ15N values in snowy egret nestlings with dietary mercury

exposure: a marker for increased protein degradation. Environmental Science and

Technology 39: 4226-4233.

Sokal, R.R., and Rohlf, F.J. 1995. Biometry. 3rd edition. W.H. Freeman and Company,

New York, 887 pp.

Sotiropoulos, M.A., Tonn, W.M., and Wassenaar, L.I. 2004. Effects of lipid extraction on

stable carbon and nitrogen isotope analyses of fish tissues: potential consequences

for food web studies. Ecology of Freshwater Fish 13: 155-160.

Spies, R.B., Kruger, H., Ireland, R., and Rice, D.W.Jr. 1989. Stable isotope ratios and

contaminant concentrations in a sewage distorted food web. Marine Ecology

Progress Series 54: 157-170.

Steele, K.W., and Daniel, R.M. 1978. Fractionation of nitrogen isotopes by animals: a

further complication to the use of variations in the natural abundance of 15N for

tracer studies. Journal of Agricultural Science 90: 7-9.

Stewart, A.R., Luoma, S.N., Schlekat, C.E., Doblin, M.A., and Hieb, K.A. 2004. Food

web pathway determines how selenium affects aquatic ecosystems: A San

Francisco Bay case study. Environmental Science and Technology 38: 4519-4526.

Suzuki, K.W., Kasai, A., Nakayama, K., and Tanaka, M. 2005. Differential isotopic

enrichment and half-life among tissues in Japanese temperate bass (Lateolabrax

japonicus) juveniles: implications for analyzing migration. Canadian Journal of

Fisheries and Aquatic Sciences 62: 671-678.

Swanson, H.K., Johnston, T.A., Leggett, W.C., Bodaly, R.A., Doucett, R.R., and Cunjak,

R.A. 2003. Trophic positions and mercury bioaccumulation in rainbow smelt

(Osmerus mordax) and native forage fishes in Northwestern Ontario lakes.

Ecosystems 6: 289-299.

Tanabe, K., and Namba, T. 2005. Omnivory creates chaos in simple food web models.

Ecology 86: 3411-3414.

Thomann, R.V., and Connolly, J.P. 1984. Model of PCB in the Lake Michigan lake trout

food chain. Environmental Science and Technology 18: 65-71.

Thompson, D.R., Bearhop, S., Speakman, J.R., and Furness, R.W. 1998. Feathers as a

means of monitoring mercury in seabirds: insights from stable isotope analysis.

Environmental Pollution 101: 193-200.

Tieszen, L.L., Boutton, T.W., Tesdahl, K.G., and Slade, N.A. 1983. Fractionation and

turnover of stable carbon isotopes in animal tissues: implications for δ13C analysis

of diet. Oecologia 57: 32-37.

Vanderklift, M.A., and Ponsard, S. 2003. Sources of variation in consumer-diet δ15N

enrichment: a meta analysis. Oecologia 136: 169-182.

Vander Zanden, M.J., and Rasmussen, J.B. 1996. A trophic position model of pelagic

food webs: impact on contaminant bioaccumulation in lake trout. Ecological

Monographs 66: 451-477.

Vander Zanden, M.J., and Rasmussen, J.B. 1999. Primary consumer δ13C and δ15N and

the trophic position of aquatic consumers. Ecology 80: 1395-1404.

Vander Zanden, M.J., and Rasmussen, J.B. 2001. Variation in δ15N and δ13C trophic

fractionation: implications for aquatic food web studies. Limnology and

Oceanography 46: 2061-2066.

Vander Zanden, M.J., and Vadeboncoeur, Y. 2002. Fishes as integrators of benthic and

pelagic food webs in lakes. Ecology 83: 2152-2161.

Wayland, M., and Hobson, K.A. 2001. Stable carbon, nitrogen, and sulfur isotope ratios

in riparian food webs on rivers receiving sewage and pulp mill effluents.

Canadian Journal of Fisheries and Aquatic Sciences 79: 5-15.

Werner, R.A., and Brand, W.A. 2001. Referencing strategies and techniques in stable

isotope ratio analysis. Rapid Communications in Mass Spectrometry 15: 501-519.

2.6 Tables and Captions

Table 2.1 Possible choices for diet-tissue 15N fractionation (∆15N) for carnivorous fishes when constructing food webs without a priori knowledge of ∆15N from captive studies

Grouping Value (‰) Reference All taxa 3.4 Minagawa & Wada 1984 2.9 Vander Zanden & Rasmussen 2001 3.4 Post 2002 2.0 McCutchan et al. 2003 2.5 Vanderklift & Ponsard 2003 Carnivores 3.2 Vander Zanden & Rasmussen 2001 2.9 McCutchan et al. 2003 2.7 Vanderklift & Ponsard 2003 Organisms that excrete ammonia 2.3 McCutchan et al. 2003 2.0 Vanderklift & Ponsard 2003

Table 2.2 Features of different tissues used for stable isotope analysis in ecotoxicology studies

Turnover Non-lethal Tissue rate sampling? Heterogeneity Other issues

Whole body intermediate No High

White muscle intermediate No* Low difficult to remove from small individuals

Liver rapid No High

Fin intermediate Yes Unknown small tissue volume; other measurements impossible

Feather slow Yes High accumulate heavy metals

Hair slow Yes High

Whole blood intermediate Yes Low

Blood plasma rapid Yes Low small tissue volume; other measurements unlikely

Blood cells intermediate Yes Low small tissue volume; other measurements unlikely

*non-lethal "muscle plugs" possible in larger fish

2.7 Figures and Captions

carbon source 1 (e.g. pelagic) [contaminant]

carbon source 2 (e.g. benthic)

δ15N

B carbon source 1 (e.g. pelagic) [contaminant]

carbon source 2 (e.g. benthic)

δ15N

Figure 2.1 Theoretical biomagnification of contaminants in contrasting food webs with similar baseline and different slopes (A) and similar slopes and different baselines (B).

Animals deriving greater than 70% of their biomass from carbon source 1 or carbon source 2 were separated using δ13C.

10 A consumer source 3 8 source 2 N 6 15 δ 4

2 source 1 0 -30-29-28-27-26-25-24-23-22-21-20

13 δ C

8 consumer source 3 7 B 6 5 source 4

15 4 δ source 1 3 2 source 2 1 0 -30 -29 -28 -27 -26 -25 -24 -23 -22 -21 -20 δ13C

Figure 2.2 Hypothetical consumer stable carbon (δ13C) and nitrogen (δ15N) isotopic

ratios (solid circles), along with three (A) and four (B) potential food sources, and back-

calculated dietary mixtures after correcting for diet-tissue 13C and 15N fractionation after

Vander Zanden and Rasmussen (2001, +’s), Post (2002, open diamonds), and McCutchan

et al. (2003, x’s). Note that in figure A, only when using fractionation estimates from

McCutchan et al. (2003) are feasible source contributions obtained, while in scenario B,

using the same fractionation estimates places the mixture outside of feasible source

contributions.

Chapter 3.0 An Elemental and Stable Isotope Assessment of Water Strider

Feeding Ecology and Lipid Dynamics: Synthesis of Lab and Field Studies

Timothy D. Jardine, Karen A. Kidd, John T. Polhemus, and Richard A. Cunjak

Published in Freshwater Biology 2008, Vol. 53: 2192-2205.

Abstract

Despite the ubiquity and abundance of water striders (Hemiptera: Gerridae) in

temperate streams and rivers and their potential usefulness as sentinels in contaminant

studies, little is known about their feeding ecology and lipid dynamics. In this study

stable isotopes of carbon (δ13C) and nitrogen (δ15N) and elemental carbon to nitrogen

ratios (C/N) were used to assess dietary habits and lipid content, respectively, for water striders. To determine diet-tissue fractionation factors, nymphs of the most common species in New Brunswick, Canada, Aquarius remigis, were reared in the lab for 73 days

and exhibited rapid isotopic turnover in response to a switch in diet (C half-life = 1.5

days, N half-life = 7.8 days). Their lipid content increased towards the end of the growing

season and resulted in lower δ13C values. Diet-tissue fractionation factors were

established after correction of δ13C data for the confounding effect of de novo lipid

13 13 15 synthesis (strider δ Cadj – diet δ C = 0.1‰, strider– diet δ N = 2.6‰). To determine

spatial and temporal patterns in water strider feeding ecology, aquatic vegetation and striders were collected from 45 streams in September and October from 2004-2007 and

from four index streams sampled repeatedly over the growing season in 2007 in New

Brunswick, Canada, and analysed for δ13C and δ15N. Water striders from the majority of

sites (83%) had less than 50% contribution of aquatic carbon to their diets but showed a gradual increase in the contribution of this carbon source to their diet with increasing stream size. Over the growing season, striders collected from the index streams varied in

13 15 13 their δ Cadj by 1.8 to 3.9‰, δ N by 1.4 to 2.8‰, and C/N by 1.5 to 2.4. Low δ Cadj variation in striders at three of the four index sites (average range within sites = 1.9‰) was likely due to the lack of difference in δ13C amongst carbon sources (aquatic and terrestrial vegetation) at those sites which may have masked temporal changes in diet.

Together these data indicate that striders exhibit a strong connection to terrestrial carbon sources, making them important users of energy subsidies to streams from the surrounding catchment. However, this dependence on terrestrial organic matter may limit their utility as indicators of contamination by heavy metals and other pollutants for aquatic food webs that are dependent on in-stream production.

3.1 Introduction

The ability to study the ecology of wild animal populations has been expanded greatly in the last 25 years through the use of stable isotope analysis (SIA). SIA has improved understanding of energy pathways (e.g. Hecky and Hesslein 1995), animal movement patterns (Hobson 1999), effects of invasive species (e.g. Vander Zanden et al.

1999) and contaminant biomagnification (e.g. Cabana and Rasmussen 1994). Accurate measurements of food source pathways are important in conservation and management

(Araujo-Lima et al. 1986) and in understanding larger patterns in food webs (Post et al.

2000). Stable carbon isotope ratios (13C/12C or δ13C) typically reveal the photosynthetic pathways supporting food webs (Hecky and Hesslein 1995) because they show little

change during food web transfer and sources of primary production often differ in their

δ13C, while stable nitrogen isotope ratios (15N/14N or δ15N) are indicators of trophic

position, as 15N becomes enriched by an average of approximately 3.4‰ with each step up a food chain (Post 2002).

Over time, the ability to use SIA to answer ecological questions has improved through the use of rearing experiments that have established diet-tissue fractionation

(Post 2002) and elemental turnover rates (Dalerum and Angerbjorn 2005) for the target species and tissues of interest (Hobson and Clark 1992), the advent of mathematically- sound mixing models to assess diet (Phillips and Gregg 2001), and the means to account for the effects of endogenous lipid synthesis on tissue δ13C values (Kiljunen et al. 2006).

However, many of the laboratory rearing experiments were aimed at understanding

isotopic fractionation and turnover in birds (e.g. Hobson and Clark 1992) and fishes (e.g.

Suzuki et al. 2005), with fewer studies on insects (e.g. Ostrom et al. 1997). Furthermore,

the uncertainty surrounding literature-based estimates of diet-tissue fractionation can

often impair precise estimates of diet for field populations for which no diet-tissue

fractionation estimate is available (Jardine at al. 2006). This necessitates the use of

rearing experiments to establish diet-tissue fractionation in cases where there is an

interest in inferring field diets using stable isotopes.

In addition to its traditional application of answering ecological questions, SIA

has proven useful as a tool to understand the nutritional state of wild animal populations,

including the recycling of proteins during starvation (Hobson et al. 1993) and the

synthesis and storage of lipids (Cherel et al. 2005). Elemental carbon to nitrogen ratios

(C/N) are indicators of lipid content of an organism (McConnaughey and McRoy 1979),

with higher C/N values associated with higher % lipid due to the presence of long chain fatty acids and a low proportion of N-bearing proteins (Campbell 1995). Given that C/N data is often provided alongside δ13C and δ15N, a simultaneous assessment of diet and

lipid content can be conducted (Cherel et al. 2005).

Water striders (Hemiptera: Gerridae) are predaceous insects that are ubiquitous

and abundant on lakes and slow-moving sections of streams and rivers around the world

(Spence and Andersen 1994). In temperate regions of North America, adults of one

common species (Aquarius remigis) overwinter and return to streams following ice-out

(Fairbairn 1985). They breed and then senesce in the spring, producing a new generation of nymphs by mid-summer that store lipids in late summer after their development to

adulthood (Lee et al. 1975). While it is generally known that water striders feed mainly

upon terrestrial and aquatic insects trapped in the surface film of the streams (Matthey

1974, MacLean 1990), little is known about the relative contributions of these two food

source pathways. Striders can make up a large proportion of the biomass of insects in

many stretches of small rivers and streams (Svensson et al. 2002). Consequently, striders

could serve as general indicators of the overall flow of energy to predatory insects and

fishes, and may also potentially be used for determining if changes occur in the

importance of different carbon sources in an upstream-downstream direction in rivers

(Vannote et al. 1980, Thorp and Delong 1994). Due to their ubiquity and abundance and

potential importance as conduits of energy into small streams, water striders therefore

offer a good model to study feeding ecology and nutritional status in wild animal

populations. Striders are also recently gaining attention as environmental sentinels for

heavy metal contamination (Jardine et al. 2005, Nummelin et al. 2007). Because much of

the contaminant burden in an animal comes from its diet (Hall et al. 1997) and because

contaminant concentrations can be affected by growth, sex, and season, it is important to

understand strider dietary and life-history patterns.

In order to investigate feeding ecology in the most common water strider species

in New Brunswick streams (A. remigis) and the relative importance of terrestrial and

aquatic carbon to its diet in time and space, a laboratory diet-switch experiment was conducted to measure diet-tissue isotope fractionation and turnover rates. Striders and

potential food sources were then sampled from four NB stream sites over the growing

season to determine the temporal changes in isotope ratios and C/N in field populations.

Adult gerrids in temperate climates typically have one to three generations per year

(Galbraith and Fernando 1977, Fairbairn 1985), and they accrue lipids over the course of

the growing season prior to overwintering (Lee et al. 1975). It was therefore predicted

that C/N, a proxy for lipid content (McConnaughey and McRoy 1979), would increase

from spring to fall. Finally water striders (mainly A. remigis but also the less common

Metrobates hesperius) and aquatic vegetation were sampled from a large number of

streams representing a range of sizes and SIA was used to assess food source pathways for striders in relation to the size of the system in which they were living to test if organic

carbon sources to consumers vary longitudinally in the river network (Vannote et al.

1980). These studies were conducted to better understand the feeding ecology of the water strider and illustrate the value in pairing field and laboratory observations to better understand ecological and nutritional processes in a predatory aquatic invertebrate.

3.2 Methods

A rearing experiment was conducted in 2006 by capturing water strider nymphs from a New Brunswick stream (Cow Pasture Brook, N 45.95 W 66.23), placing them in two plastic tanks (volume = 48L, initial density ≈ 40 individuals per tank) with standing water and feeding them crickets obtained from a single batch at a commercial pet food store. Striders were held on the diet for approximately 72 days, with samples (n = 2) removed from each of the two tanks on days 0, 1, 2, 4, 8, 16, 23, 31, 41, 48, 54, and 61, and from one tank on days 68 and 73. The amount of food added to the tanks was adjusted over the course of the experiment in response to decreased strider densities, with approximately one cricket added for every two striders present. Striders from the source stream (Cow Pasture Brook) were also re-sampled on day 51 of the experiment to compare isotope ratios with changes observed as a result of the diet-switch in the laboratory.

In the field, water striders were collected using hand nets from 81 stream sites in

New Brunswick, Canada in the months of September and October from 2004-2007.

Thirty-four sites were sampled in 2004, nine in 2005, 24 in 2006 (with a return visit to one site), and 26 in 2007 (including return visits to 11 sites). When sites were sampled in more than one year, data were combined to yield one point for each site assuming no difference between years. Streams were assigned orders manually with maps at the

1:150,000 scale. Sites ranged from 1st to 6th order systems, varied in width from one to

>50 m, and had limited or no catchment disturbance from forestry, agriculture or urbanization. An additional four sites (Corbett Brook, N 45.92 W 66.64, English Brook

N 46.43 W 66.60, McKenzie Brook N 46.22 W 66.53, and Parks Brook N 45.46 W

66.35) were sampled approximately bi-weekly from May to October 2007.

A. remigis was the most commonly captured water strider, appearing at 72 of the

81 sites (including the four index sites). At the index sites, in two instances (July 3rd at

Parks Brook, July 17th at Corbett Brook) only nymphs were in sufficient abundance at the time of collections. The other, smaller-bodied, gerrid species (M. hesperius) occurred at

16 sites (including Parks Brook, one of the index sites). The two species were found

together at only seven sites. At each site where striders were captured, multiple samples

of two individuals each were submitted for SIA. In 2004 and 2005 sexes were not

identified, while in 2006 and 2007 sexes were analyzed separately.

To characterize the different carbon sources in these streams, aquatic mosses

(Fontinalis sp.), macrophytes (e.g. Potomageton sp.), and filamentous algae (e.g.

Cladophora sp.) were collected randomly by hand from each site in 2004. In the years

2005 to 2007 aquatic vegetation was sampled by scrubbing rocks with a toothbrush to collect biofilm (n = 3 samples of a minimum of three rocks per sample chosen randomly within the boundaries of the site). Biofilm is an amalgamation of different types of aquatic vegetation (diatoms, moss, filamentous algae) as well as fungi and bacteria

(Battin et al. 2003), and thus closely represents in-stream primary production. Any carbon fixed within the stream was therefore considered aquatic carbon. The pH of the streams ranged from 4.7 to 8.1, with the majority of values between 6.5 and 7.5. Most streams were therefore circumneutral or acidic and calcareous soils are uncommon in

New Brunswick, hence biofilm samples were not acidified to remove carbonates that can confound δ13C.

All samples for SIA were oven-dried @ 60ºC for 48h and ground to a powder

using a mortar and pestle, or ball-mill grinder. Samples were analyzed for stable carbon

and nitrogen isotope ratios by combustion in a Carlo Erba NC2500 Elemental Analyzer and delivery to a Finnigan Mat Delta Plus mass spectrometer (Thermo Finnigan, Bremen,

Germany) via continuous flow. Data are reported as delta values relative to international standards Peedee Belemnite Carbonate (PDB) and atmospheric nitrogen (AIR) and calibrated using International Atomic Energy Agency standards CH6 (-10.4‰), CH7 (-

31.8‰), N1 (0.4‰) and N2 (20.3‰). A single water strider sample (working laboratory standard) was analyzed every time samples were run to assess precision across analytical runs, and yielded values of δ13C = -27.53 ± 0.20‰ (unless otherwise indicated, data are

shown as mean ± S.D.) and δ15N = 4.21 ± 0.65‰ (n = 11). The poorer precision in δ15N was due to a single analysis of this standard, and several samples analyzed in that particular run were repeated to ensure values were comparable with other analytical runs.

A commercially available standard (acetanilide, Elemental Microanalysis, Ltd.) analyzed alongside the samples yielded values of δ13C = -33.60 ± 0.12‰ and δ15N = -3.06 ±

0.25‰ (n = 59).

Given that high lipid levels (as indicated by high C/N) can drive δ13C in a

negative direction (McConnaughey and McRoy 1979, Matthews and Mazumder 2005),

more negative δ13C was expected later in the season as gerrids accrued lipids before the

13 onset of winter. To normalize stable carbon data (denoted δ Cadj) and account for

varying lipid levels across individuals and sites, δ13C was corrected using an equation (r2

= 0.50) outlined in Logan et al. (2008):

13 13 δ Cadj = δ C – (2.08 – 1.92 * ln (C/N))

This equation was developed by lipid extracting a series of pooled A. remigis samples after their initial analysis for δ13C and δ15N. The samples were then re-analyzed

following extraction and data compared to initial C/N. Using this equation to correct for

lipids reduces within-site variability in A. remigis δ13C (Logan et al. 2008).

One-isotope, two source mixing models (Phillips and Gregg 2001) were used to

estimate the percentage of carbon from each source (terrestrial and aquatic) in the diet of striders. Mixing models compare the isotope ratio of the consumer with isotope ratios in underlying food sources. Aquatic vegetation has been shown to have highly variable

δ13C both within and across sites (Osmond et al. 1981) and overlaps the range observed

for terrestrial vegetation (France 1995). In the mixing models, analyses were run with a

δ13C value of -28.2 ± 0.9‰ S.D. (Finlay 2001) for terrestrial detritus (coarse particulate

organic matter) that agreed with an earlier estimate reported by France (1995), and a site-

specific δ13C for aquatic vegetation from measurements conducted herein. The value for

terrestrial detritus also approximates that reported for riparian arthropods by Collier et al.

(2002) and Paetzold et al. (2005). To avoid drawing false conclusions about food source

pathways where the two sources potentially overlapped, proportions of the different

dietary sources were only quantitatively assessed using data from sites where aquatic

vegetation had -30‰ > δ13C > -26‰ (i.e. values outside the likely range of terrestrial

vegetation δ13C; France 1995, Finlay 2001). This narrowed the number of sites from the

88 sampled down to a total of 45 sites. It also eliminated three of the four index sites,

with only Parks Brook consistently having aquatic vegetation that was distinct from

terrestrial organic matter. Striders with δ13C outside the δ13C range delimited by

terrestrial and aquatic vegetation (after correcting for slight diet-tissue fractionation from

the rearing experiment) were constrained to values of 0 or 100% (Vander Zanden and

Vadeboncoeur 2002).

Data were analyzed using NCSS (Kaysville, UT) and SYSTAT (version 9, SPSS,

Inc., Chicago, IL) software. For the laboratory experiment, exponential models (y = b +

ct 13 13 15 a*e ) were used to fit the isotope data (δ C, δ Cadj and δ N) vs. time, following

Hobson and Clark (1992), where t is the time (in days) since the diet switch and c is the derived constant. Half-life was calculated by the formula: half-life = ln(0.5)/c. Diet tissue fractionation was calculated by subtracting isotope ratios of crickets (δ13C = -24.1

13 15 ± 0.5‰, δ Cadj = -23.5 ± 0.2‰, δ N = 3.8 ± 0.2‰, n = 6) from that of striders.

For the field data, strider δ13C was adjusted for lipids (using C/N) and diet-tissue

fractionation (multiplying by two to represent the two trophic levels from primary producers to striders) and then used to estimate percent aquatic carbon in the diet.

Comparisons of percent aquatic carbon in the diet of striders to stream size (order) at the

45 sites sampled in the fall were made using linear regressions. To examine whether dietary habits or lipids varied significantly over time or between sexes at the four streams that were regularly sampled, data were grouped into two generations, the first that returned to stream surfaces after overwintering (“post-winter”) and the new generation that hatched in early summer (“pre-winter”). Adults sampled in the period after ice-out in the spring up to and including the first day when a new generation of nymphs was present (typically early July) were considered as the post-winter sample. Adults sampled for the remainder of the growing season (July – October) were made up of a new generation and considered pre-winter. The timing of the arrival of the new generation, and hence the delineation of the two samples, varied slightly from one stream to another

13 (see Table 3.1 for details). Comparisons of adult water strider isotope ratios (δ Cadj and

δ15N) and C/N across sampling periods were made using general linear model analysis of

variance (GLM-ANOVA) with three factors – site (random factor: Corbett Brook,

English Brook, McKenzie Brook, and Parks Brook), sex (fixed factor: male and female),

and generation (fixed factor: post- and pre-winter). Paired comparisons of differences among levels of fixed factors were made using the Bonferroni test. Finally, average values for males and females were compared to those of nymphs for those times that they co-occurred (mainly summer) with two factors (stage – fixed, and site - random) in a

GLM-ANOVA. Interactions and main effects were considered significant at α = 0.05 for all tests.

3.3 Results

In the lab, A. remigis nymphs grew rapidly on the cricket diet from a wet weight of

10.8 ± 5.2 mg (3.0 ± 1.5 mg dry weight (d.w.)) to a final wet weight of 49.3 ± 2.7 mg

(21.2 ± 1.9 mg d.w.) as adults on day 73 of the experiment (Fig. 3.1a). As is common in water striders (Fairbairn 2005), females reached greater body sizes than males and had

37% higher wet weights (asymptotic values) (Fig. 3.1a). Striders showed an abrupt increase in C/N between days 31 and 41 (Day 31 C/N = 3.84 ± 0.03; Day 41 C/N = 5.17

± 0.53, p = 0.002, Fig. 3.1b), corresponding to Julian Days of 217 and 227 (late July/early

August). From day 41 onward, C/N remained elevated in both males and females until

13 13 15 the end of the rearing experiment (Fig. 3.1b). Initial δ C, δ Cadj and δ N of A. remigis

upon being brought to the lab were -26.9 ± 0.3‰, -26.5 ± 0.3‰, and 3.2 ± 0.2‰,

13 13 15 respectively. Asymptotic values were -24.3‰ (δ C), -23.4‰ (δ Cadj) and 6.4‰ (δ N)

and achieved (>99% of asymptote) on days four, 16, and 48, respectively (Fig. 3.2). The

13 best-fit exponential model for isotopic change with time obtained when using δ Cadj captured more variation (r2 = 0.78) compared with using unadjusted δ13C (r2 = 0.68) (Fig.

3.2). By subtracting isotope ratios for crickets from asymptote values for A. remigis,

diet-tissue fractionation for this species was calculated as -0.2‰ for δ13C, 0.1‰ for

13 15 δ Cadj and 2.6‰ for δ N. A. remigis showed a rapid turnover of carbon (half-life = 1.1

13 13 days modeled using δ C and half-life = 1.5 days modeled using δ Cadj, Fig. 3.2) and a

relatively fast turnover of nitrogen (half-life = 7.8 days, Fig. 3.2).

In the field, aquatic vegetation δ13C varied widely among sites; mean δ13C values

ranged from -39.1‰ to -18.2‰ and increased with increasing stream order (δ13C =

2.2*Order – 36.0, r2 = 0.39, p < 0.0001). The range of δ13C of aquatic vegetation enveloped the δ13C for terrestrial vegetation estimated from the literature (~ -28‰,

France 1995, Finlay 2001).

The percentage of aquatic carbon in the diet of A. remigis increased with increasing stream size (p < 0.001, Fig. 3.3). Striders rarely showed any evidence of even minimal consumption of aquatic insects in streams of order three and smaller. In 4th

order streams, the entire range of aquatic carbon contribution (0% to 100%) was

observed, while in the largest systems (5th and 6th order) striders had a mix of aquatic and terrestrial carbon in their diet (range = 25-89% aquatic carbon). M. hesperius was only present in larger streams (≥ 4th order) and appeared to have similar proportions of aquatic

carbon as A. remigis at those sites (Fig. 3.3).

Stable isotope ratios and C/N of striders varied among sexes and generations

within the four streams (Table 3.1). Striders collected from Corbett Brook, English

13 Brook, and McKenzie Brook showed little change in δ Cadj over the course of the

growing season, with a minimum value of -27.6‰ and a maximum value of -25.1‰.

Because there was no difference between δ13C of aquatic vegetation and the assumed

terrestrial end-member at Corbett Brook, English Brook, and McKenzie Brook, % aquatic

carbon in the diet of striders was only estimated from Parks Brook. At this site, the

13 dietary source of C for striders varied over time with a decrease in δ Cadj in males,

females and nymphs occurring in mid-summer (Table 3.1) due to increased reliance on

aquatic carbon (Fig. 3.4). In the spring and late summer, striders at Parks Brook had less than 40% of their carbon coming from aquatic sources (Fig. 3.4). For δ15N, there were no

consistent temporal trends within the four sites, with δ15N peaks occurring randomly

throughout the growing season (e.g. Corbett Brook June 11th and July 17th males, English

Brook July 4th males, McKenzie Brook July 4th females, Parks Brook September 17th males). There were no consistent differences in δ15N between males and females within streams, but nymphs had consistently lower δ15N than adults at all sites (average

difference = 0.8 ± 0.7‰, range = -0.1 to 2.1‰, p = 0.023, Table 3.1). At all sites, higher

C/N were observed in late summer and fall (C/N > 5) compared with spring (C/N ≈ 4)

(Table 3.1), and this was similar in timing and magnitude to the increase in C/N observed in the experimental population (Fig. 3.1b).

Across streams, there were significant differences in stable isotope ratios and

C/N. No interactions, including three way interactions (sex x generation x site), were

13 significant (p > 0.05). For δ Cadj, there were significant differences among sites (p <

0.001) but no differences among sexes (p = 0.056) or generations (p = 0.913). There

13 were no differences in δ Cadj between adults and nymphs (p = 0.298, Table 3.1). Strider

δ15N values were also significantly different across sites (p < 0.001) while sexes (p =

0.053) and generations (p = 0.916) did not differ in their δ15N. For C/N, only generation was significant (p = 0.009), with the 2nd generation (pre-winter) having higher C/N than overwintered adults (post-winter). There were no differences between sexes (p = 0.273) or sites (p = 0.475) in C/N. Adults did not have significantly higher C/N than nymphs (p

= 0.055).

3.4 Discussion

By combining a lab diet-switch experiment with field collections, stable isotopes and C/N were used to gain insights into water strider feeding ecology and lipid dynamics.

Data indicate that striders have a strong connection to terrestrial carbon sources, particularly A. remigis inhabiting small streams. They also suggest that strider nymphs exhibit rapid elemental turnover during development, and that a pronounced increase in lipid synthesis occurs in late summer in advance of the over-wintering period.

Estimates of turnover rate for nitrogen (half-life = 7.8 days) and carbon (half-life ≈

2 days) were faster than has been previously observed in rapidly-growing ectotherms

(Suzuki et al. 2005) and similar to metabolically active tissues, such as liver, in endotherms with high energy demands (Tieszen et al. 1983). The rapid turnover observed for striders may be due to a fast metabolism and resultant high energy demands characteristic of insects (Altman and Dittmer 1968). The difference in half-life between carbon and nitrogen may be related to differential metabolic needs (e.g. energy vs. growth) during nymph development. Ostrom et al. (1997) found rapid carbon turnover in beetle tissues (Harmonia variegate Goeze), with 75% change occurring in six days

following a diet switch. In contrast, nitrogen turnover in the same organism was slower,

with 75% change occurring in 21 days. Faster turnover of carbon than nitrogen has been shown in some other diet-switch experiments with stable isotopes (Hobson and Bairlein

2003, Suzuki et al. 2005), but in most cases turnover rates of these two elements have been similar (Bearhop et al. 2002, Haramis et al. 2003, Pearson et al. 2003, Evans-Ogden et al. 2004, Doi et al. 2007). A faster turnover of carbon than nitrogen may be due to the rapid uptake and breakdown of labile energy sources (e.g. sugars and lipids) during growth, as well as whole-body changes in biochemical composition that occur during moulting (Doi et al. 2007). Striders grew from 2nd and 3rd instars to adults during the course of the rearing experiment and hence would have moulted more than once during this period.

There was a better fit of the isotopic change model to the carbon-13 data when

13 13 using δ Cadj as opposed to δ C, suggesting that lipids could potentially interfere with

interpretation of δ13C of animal populations in the field. The decline in δ13C towards the end of the experiment as striders began synthesizing lipids (Fig. 3.2) may be similar to that observed by Gratton and Forbes (2006) for tissues of female ladybeetles (H.

axyridis). In that study, ladybeetles were switched to a corn based diet that was enriched in 13C, and after an initial increase in δ13C, their tissues became more 13C-depleted near the end of the study. However, because no elemental % data (i.e. C/N) were available,

δ13C remained uncorrected for lipids and the interpretation of the late decline in δ13C was

left to speculation (Gratton and Forbes 2006). The abrupt increase in C/N in both lab-

reared striders and in field populations suggests a switch to lipid synthesis prior to the

overwintering period. Given that striders effectively hibernate for ~six months

(November to April) in northern North America (Fairbairn 1985), the storage of

sufficient lipid reserves may be critical to survival. Similar suggestions of the importance of lipids have been made for other ectotherms in northern climates (e.g. fishes, Biro et al.

2004, Hurst and Conover 2003).

Estimates of diet-tissue 13C and 15N fractionation for striders in the current study

are consistent with ranges reported previously for a variety of species (Post 2002,

McCutchan et al. 2003). McCutchan et al. (2003) recently suggested that fluid feeders

may exhibit lower 15N fractionation than other taxa, however the value observed for

fluid-feeding water striders (2.6‰) is well within the range reported for most animals

(Post 2002, McCutchan et al. 2003, Vanderklift and Ponsard 2003). By obtaining a diet-

tissue fractionation 13C estimate (∆δ13C) specific to A. remigis, the use of literature

fractionation estimates were avoided that, because they summarize data for a large

number of species, contain a large amount of uncertainty in their application and error

(e.g. ∆δ13C = 0.4 ± 1.3‰ S.D., Post 2002, McCutchan et al. 2003). This improved confidence in mixing model estimates for % aquatic energy in the diet of striders.

Isotope ratios of aquatic vegetation and their importance to the diet of striders increased from small headwater streams to larger, wider stretches of river. The ability to detect changes in the diet of striders, however, was limited by the necessity of eliminating

~half of the sites because δ13C of aquatic vegetation was similar to that of terrestrial

organic matter. Conclusions, therefore, apply only to those sites (45 of 88) where aquatic

and terrestrial vegetation had different δ13C values. Aquatic vegetation δ13C generally

increased with increasing stream order, consistent with observations by Finlay (2001)

who found a significant relationship between aquatic vegetation δ13C and drainage basin

size. This observed change is likely a consequence of increased CO2 availability in CO2-

saturated headwater streams, which leads to greater discrimination against 13C during

photosynthesis and lower δ13C in aquatic algae in smaller, more turbulent systems (Finlay

2004). A. remigis showed a limited increase in % aquatic carbon in a downstream direction, suggesting a continued reliance on terrestrial vegetation even in medium sized systems (stream orders 3 and 4). Only at seven of 42 sites did A. remigis show a dominant contribution (>50%) from aquatic food sources in the diet. This channeling of terrestrial production into a supposed aquatic consumer is another example of a reciprocal subsidy (Baxter et al. 2005), wherein seemingly distinct environments are connected by energy and material flow across boundaries.

A high proportion of terrestrial carbon in the diet of A. remigis estimated from mixing models supports field observations of the species’ habitat associations. A. remigis is often found in riparian edges, even in larger systems (T.D.J., pers. obs.), which may result in considerable consumption of terrestrial insects that fall onto the water surface.

The majority of A. remigis growth occurs later in the summer, when terrestrial insect supply to streams is typically maximized relative to aquatic insect emergence (Nakano and Murakami 2001). A. remigis nymphs typically appear on stream surfaces in July and grow rapidly. Early in their development, these nymphs may feed on various microbes from the surface film of the water as well as catching the end of the peak in insect emergence. The trophic level of nymphs is low relative to adults as indicated by their lower δ15N. Estimates of turnover rates reported here suggest that isotope ratios in field

populations should respond rapidly to changes in availability of different prey types

provided those prey types have distinct isotope ratios. At Parks Brook where % aquatic

carbon in the diet was estimated over the course of the growing season, the peak in %

aquatic carbon (~60% for males, females, and nymphs) was observed in mid-July and the

importance of this carbon source declined thereafter.

A second possibility for explaining the strong connection between A. remigis and

terrestrial carbon may be because the striders rely on aquatic insects that directly process

leaf litter and its associated microbes (Wallace et al. 1997). Doucett et al. (1996) found a strong terrestrial δ13C signal in muscle tissue of juvenile Atlantic salmon (Salmo salar) in the upper reach of a New Brunswick stream that likely originated from the consumption of abundant “shredder” taxa such as the stoneflies Pteronarcyidae. Perry et al. (2003)

found strong evidence for terrestrial carbon in the diet of juvenile Chinook salmon

(Oncorhynchus tshawytscha) in five headwater tributaries of the Yukon River that may

have resulted from consumption of a shredder Tipulidae in those streams. Given the ubiquity and abundance of A. remigis in small streams (1st to 4th order, Svensson et al.

2002), the species may be a conduit for the incorporation and possible recycling of

terrestrial energy in these types of systems (Vannote et al. 1980).

M. hesperius was only present in lower reaches of rivers (Order ≥ 4) suggesting

that this species may use more energy derived from aquatic food sources. However it

was difficult to estimate source proportions using mixing models for this species; M.

hesperius is far less common than A. remigis in New Brunswick streams, species-specific

diet-tissue fractionation or lipid correction estimates were lacking, and aquatic vegetation

δ13C was similar to terrestrial vegetation at many sites where M. hesperius occurred.

Only at eight sites were estimates of source proportions for this species obtained and

values of % aquatic carbon in their diets ranged from 25% to 100% (4% to 78% at the

index site Parks Brook). M. hesperius rarely occurred in sympatry with A. remigis and was absent from all streams that were < ~8 m in width. Kittle (1977) similarly reported that M. hesperius occurred in large and medium-sized streams and was not typically found in association with other gerrids; where present, it was most commonly located above and below riffles (i.e. in pools). Given that algae growing in pools typically exhibit enriched δ13C (Finlay et al. 2002), consumption of emerging insects from these habitats may explain the relatively enriched δ13C in M. hesperius at some of the sites where the species did occur. At Parks Brook, A. remigis and M. hesperius appeared to feed on similar resources suggesting that the lack of occurrence of M. hesperius in smaller streams may not be driven by diet limitations but other factors such as stream turbulence or predation.

The extensive use of terrestrial carbon by A. remigis has implications for studies that seek to use this species to identify aquatic systems with high contaminant loads

(Jardine et al. 2005, Nummelin et al. 2007). Given the importance of diet as the primary route of exposure of organisms to contaminants such as mercury (Hall et al. 1997), a disconnection between striders and emerging aquatic insects as a diet and contaminant source would limit their usefulness for monitoring pollutants in aquatic systems if the pollutants were originating in the aquatic environment. As air breathers that are not immersed in the aquatic environment, striders also lack a secondary source of exposure to contaminants (e.g. waterborne exposure to fishes) leaving diet as their only contaminant source. If the diet is composed primarily of terrestrial sources, strider contaminant concentrations will be more representative of processes occurring in the terrestrial environment, such as atmospheric deposition.

Seasonal variation in strider dietary habits could have implications for the

understanding of their long-term diet as indicated by SIA. Short-term use of aquatic

carbon in July by A. remigis at Parks Brook occurred prior to the onset of larger-scale

one-time sampling at other sites. The majority of sites were sampled in August and

September, a time that corresponds to when striders at Parks Brook were back to feeding

on terrestrial carbon after short term reliance (> 50%) on aquatic sources. However,

because the bulk of the diet is made up of terrestrial sources, as at Parks Brook where %

terrestrial carbon averaged 73% when all sexes and dates were considered, sampling in

late summer will indeed be appropriate as an indicator of long-term diet.

The large amount of error surrounding estimates of % aquatic carbon in the diet of

A. remigis is largely due to the variability in aquatic vegetation δ13C within sites. Due to heterogeneity in parameters such as water velocity and light availability which ultimately

13 affect CO2 delivery (Finlay 2004) within stream reaches, aquatic vegetation δ C can

vary widely (France 1995). Sampling rigor in the current study was likely not

appropriate to fully capture all variability within a site given small sample sizes of

aquatic vegetation (n = 1-3 pooled samples per site) and possible contamination of

biofilm samples by terrestrial detritus (Hamilton et al. 2005). However, the use of a

laboratory diet-switch experiment allowed a better estimate for diet-tissue fractionation

for the species of interest, and hence more valid estimates of proportions of different

dietary sources in the field populations. It is recommended that future studies with stable

isotopes perform simple laboratory rearing trials to gain this important information and

better assess diet in wild animal populations. It is also recommended that close attention

be paid to the information provided by elemental C/N as both an indicator of the

confounding influence of lipids on δ13C (McConnaughey and McRoy 1979, Logan et al,

2008) and, perhaps more critically, as an index of the lipid content of the organism

(Cherel et al. 2005).

3.5 Acknowledgements

The authors wish to thank T. Arciszewski, P. Emerson, S. Fraser, L. Giardi, K.

Lippert, A. Townsend, J. Penny, T. McMullen, C. Blanar, C. Poole, R. Keeler, P. Brett,

T. Barrett, R. Engelbertink, A. Fraser, M. Sullivan, S. McWilliam, N. Swain, and M.

Sabean for their assistance in field sampling, D. Fairbairn and A. Sih for advice on water strider ecology, and A. McGeachy, M. Savoie, and C. Paton at the Stable Isotopes in

Nature Laboratory (University of New Brunswick) for conducting the isotope analyses.

L. Jardine, N. Burgess, K. Munkittrick, and two anonymous reviewers provided comments on an earlier draft of this manuscript. The Collaborative Mercury Research

Network (COMERN), NSERC Discovery, Canada Graduate Scholarship and Canada

Research Chair programmes, the New Brunswick Environmental Trust Fund, and the

New Brunswick Wildlife Trust Fund provided funding for this research.

3.6 References

Altman, P. L., and Dittmer, D.S. 1968. Metabolism. Federation of American Societies for

Experimental Biology, Bethesda, MD.

Araujo-Lima, C.A.R.M., Forsberg, B.R., Victoria, R., and Martinelli, L. 1986. Energy

sources for detritivorous fishes in the Amazon. Science 234: 1256-1258.

Battin, T.J., Kaplan, L.A., Newbold, J.D., and Hansen, C.M.E. 2003. Contributions of

microbial biofilms to ecosystem processes in stream mesocosms. Nature 426: 439-

442.

Baxter, C.V., Fausch, K.D., and Saunders, W.C. 2005. Tangled webs: reciprocal flows of

invertebrate prey link streams and riparian zones. Freshwater Biology 50: 201-220.

Bearhop, S., Waldron, S., Votier, S.C., and Furness, R.W. 2002. Factors that influence

assimilation rates and fractionation of nitrogen and carbon stable isotopes in avian

blood and feathers. Physiological and Biochemical Zoology 75: 451-458.

Biro, P.A., Morton, A.E., Post, J.R., and Parkinson, E.A. 2004. Over-winter lipid

depletion and mortality of age-0 rainbow trout (Oncorhynchus mykiss). Canadian

Journal of Fisheries and Aquatic Sciences 61: 1513-1519.

Cabana, G., and Rasmussen, J.B. 1994. Modelling food chain structure and contaminant

bioaccumulation using stable nitrogen isotopes. Nature 372: 255-257.

Campbell, M.K. 1995. Biochemistry. 2nd Ed. Saunders College Publishing, Philadelphia.

657 pp.

Cherel, Y., Hobson, K.A., Bailleul, F., and Groscolas, R. 2005. Nutrition, physiology,

and stable isotopes: new information from fasting and molting penguins. Ecology

86: 2881-2888.

Collier, K.J., Bury, S., and Gibbs, M. 2002. A stable isotope study of linkages between

stream and terrestrial food webs through spider predation. Freshwater Biology 47:

1651-1659.

Dalerum, F., and Angerbjorn, A. 2005. Resolving temporal variation in vertebrate diets

using naturally occurring stable isotopes. Oecologia 144: 647-658.

Doi, H., Kikuchi, E., Takagi, S., and Shikano, S. 2007. Changes in carbon and nitrogen

stable isotopes pf chironomid larvae during growth, starvation and metamorphosis.

Rapid Communications in Mass Spectrometry 21: 997-1002.

Doucett, R.R., Power, G., Barton, D.R., Drimmie, R.J., and Cunjak, R.A. 1996. Stable

isotope analysis of nutrient pathways leading to Atlantic salmon. Canadian Journal

of Fisheries and Aquatic Sciences 53: 2058-2066.

Evans-Ogden, L.J., Hobson, K.A., and Lank, D.B. 2004. Blood isotopic (δ13C and δ15N)

turnover and diet-tissue fractionation factors in captive dunlin (Calidris alpina

pacifica). The Auk 121: 170-177.

Fairbairn, D.J. 1985. Comparative ecology of Gerris remigis (Hemiptera: Gerridae) in

two habitats: a paradox of habitat choice. Canadian Journal of Zoology 63: 2594-

2603.

Fairbairn, D.J. 2005. Allometry for sexual size dimorphism: testing two hypotheses for

Rensch's Rule in the water strider Aquarius remigis. The American Naturalist 166

(Suppl): 69-84.

Finlay, J.C. 2001. Stable-carbon-isotope ratios of river biota: implications for energy

flow in lotic food webs. Ecology 82: 1052-1064.

Finlay, J.C. 2004. Patterns and controls of lotic algal carbon isotope ratios. Limnology

and Oceanography 49: 850-861.

Finlay, J.C., Khandwala, S., and Power, M.E. 2002. Spatial scales of carbon flow in a

river food web. Ecology 83: 1845-1859.

France, R. 1995. Critical examination of stable isotope analysis as a means for tracing

carbon pathways in stream ecosystems. Canadian Journal of Fisheries and Aquatic

Sciences 52: 651-656.

Galbraith, D.F., and Fernando, C.H. 1977. The life history of Gerris remigis

(Heteroptera: Gerridae) in a small stream in southern Ontario. Canadian

Entomologist 109: 221-228.

Gratton, C., and Forbes, A.E. 2006. Changes in δ13C stable isotopes in multiple tissues of

insect predators fed isotopically distinct prey. Oecologia 147: 615-624.

Hall, B.D., Bodaly, R.A., Fudge, R.J.P., Rudd, J.W.M., and Rosenberg, D.M. 1997. Food

as the dominant pathway of methylmercury uptake by fish. Water, Air, and Soil

Pollution 100: 13-24.

Hamilton, S.K., Sippel, S.J., and Bunn, S.E. 2005. Separation of algae from detritus for

stable isotope or ecological stoichiometry studies using density fractionation in

colloidal silica. Limnology and Oceanography: Methods 3: 149-157.

Haramis, G.M., Jorde, D.G., Macko, S.A., and Walker, J.L. 2001. Stable-isotope analysis

of canvasback winter diet in Upper Chesapeake Bay. The Auk 118: 1008-1017.

Hecky, R.E., and Hesslein, R.H. 1995. Contributions of benthic algae to lake food webs

as revealed by stable isotope analysis. Journal of the North American Benthological

Society 14: 631-653.

Hobson, K.A. 1999. Tracing origins and migration of wildlife using stable isotopes: a

review. Oecologia 120: 314-326.

Hobson, K.A., and Clark, R.G. 1992. Assessing avian diets using stable isotopes I:

turnover of 13C in tissues. The Condor 94: 181-188.

Hobson, K.A., Alisauskas, R.T., and Clark, R.G. 1993. Stable-nitrogen isotope

enrichment in avian tissues due to fasting and nutritional stress: implications for

isotopic analyses of diet. The Condor 95: 388-394.

Hobson, K.A., and Bairlein, F. 2003. Isotopic fractionation and turnover in captive

Garden Warblers (Sylvia borin): implications for delineating dietary and migratory

associations in wild passerines. Canadian Journal of Zoology 81: 1630-1635.

Hurst, T.P., and Conover, D.O. 2003. Seasonal and interannual variation in the allometry

of energy allocation in juvenile striped bass. Ecology 84: 3360-3369.

Jardine, T.D., Al, T.A., MacQuarrie, K.T.B., Ritchie, C.D., Arp, P.A., Maprani, A., and

Cunjak, R.A. 2005. Water striders (family Gerridae): Mercury sentinels in small

freshwater ecosystems. Environmental Pollution 134: 165-171.

Jardine, T.D., Kidd, K.A., and Fisk, A.T. 2006. Applications, considerations and sources

of uncertainty when using stable isotope analysis in ecotoxicology. Environmental

Science and Technology 40: 7501-7511.

Kiljunen, M., Grey, J., Sinisalo, T., Harrod, C., Immonen, H., and Jones, R.I. 2006. A

revised model for lipid-normalizing δ13C values from aquatic organisms, with

implications for isotope mixing models. Journal of Applied Ecology 43: 1213-1222.

Kittle, P.D. 1977. The biology of water striders (Hemiptera: Gerridae) in Northwest

Arkansas. American Midland Naturalist 97: 400-410.

Lee, R.F., Polhemus, J.T., and Cheng, L. 1975. Lipids of the water-strider Gerris remigis

Say (Heteroptera: Gerridae). Seasonal and developmental variations. Comparative

Biochemistry and Physiology 51: 451-456.

Logan, J.M., Jardine T.D., Miller, T.J., Bunn, S.E., Cunjak, R.A., and Lutcavage, M.E.

2008. Lipid corrections in carbon and nitrogen stable isotope analyses: comparison

of chemical extraction and modeling methods. Journal of Animal Ecology 77: 838-

846.

Matthews, B., and Mazumder, A. 2005. Temporal variation in body composition (C:N)

helps explain seasonal patterns of zooplankton δ13C. Freshwater Biology 50 : 502-

515.

Matthey, W. 1974. Contribution à l’écologie de Gerris remigis Say sur deux étangs des

Montagnes Rocheuses. Bulletin de la Société Suisse d’Entomologie 47: 85-95.

McConnaughey, T., and McRoy, C.P. 1979. Food-web structure and the fractionation of

carbon isotopes in the Bering Sea. Marine Biology 53: 257-262.

McCutchan, J.H.Jr., Lewis, W.M.Jr., Kendall, C., and McGrath, C.C. 2003. Variation in

trophic shift for stable isotope ratios of carbon, nitrogen, and sulfur. Oikos 102:

378-390.

McLean, E.B. 1990. Sexual dimorphism and predaceous feeding habits of the

waterstrider Gerris remigis Say (Heteroptera: Gerridae). Canadian Journal of

Zoology 68: 2688-2691.

Nakano, S., and Murakami, M. 2001. Reciprocal subsidies: Dynamic interdependence

between terrestrial and aquatic food webs. Proceedings of the National Academy of

Sciences U.S.A. 98: 166-170.

Nummelin, M., Lodenius, M., Tulisalo, E., Hirvonen, H., and Alanko, T. 2007. Predatory

insects as bioindicators of heavy metal pollution. Environmental Pollution 145:

339-347.

Osmond, C.B., Valaane, N., Haslam, S.M., Uotila, P., & Roksandic, Z. 1981

Comparisons of δ13C values in leaves of aquatic macrophytes from different

habitats in Britain and Finland; some implications for photosynthetic processes in

aquatic plants. Oecologia 50: 117-124.

Ostrom, P.H., Colunga-Garcia, M., and Gage, S.H. 1997. Establishing pathways of

energy flow for insect predators using stable isotope ratios: field and laboratory

evidence. Oecologia 109: 108-113.

Paetzold, A., Schubert, C.J., and Tockner, K. 2005. Aquatic terrestrial linkages along a

braided river: Riparian arthropods feeding on aquatic insects. Ecosystems 8: 748-

759.

Pearson, S.F., Levey, D.J., Greenberg, C.H., and Martinez del Rio, C. 2003. Effects of

elemental composition on the incorporation of dietary nitrogen and carbon isotopic

signatures in an omnivorous songbird. Oecologia 135: 516-523.

Perry, R.W., Bradford, M.J., and Grout, J.A. 2003. Effects of disturbance on contribution

of energy sources to growth of juvenile chinook salmon (Oncorhynchus

tschawytscha) in boreal streams. Canadian Journal of Fisheries and Aquatic

Sciences 60: 390-400.

Phillips, D.L., and Gregg, J.W. 2001. Uncertainty in source partitioning using stable

isotopes. Oecologia 127: 171-179.

Post, D.M. 2002. Using stable isotopes to estimate trophic position: models, methods, and

assumptions. Ecology 83: 703-718.

Post, D.M., Pace, M.L., and Hairston, N.G.Jr. 2000. Ecosystem size determines food-

chain length in lakes. Nature 405: 1047-1049.

Spence, J.R., and Andersen, N.M. 1994. Biology of water striders: interactions between

systematics and ecology. Annual Reviews in Entomology 39: 101-128.

Suzuki, K.W., Kasai, A., Nakayama, K., and Tanaka, M. 2005. Differential isotopic

enrichment and half-life among tissues in Japanese temperate bass (Lateolabrax

japonicus) juveniles: implications for analyzing migration. Canadian Journal of

Fisheries and Aquatic Sciences 62: 671-678.

Svensson, A.M., Danielsson, I., and Rydell, J. 2002. Avoidance of bats by water striders

(Aquarius najas, Hemiptera). Hydrobiologia 489: 83-90.

Thorp, J.H., and Delong, M.D. 1994. The riverine productivity model: an heuristic view

of carbon sources and organic processing in large river ecosystems. Oikos 70: 305-

308.

Tieszen, L.L., Boutton, T.W., Tesdahl, K.G., and Slade, N.A. 1983. Fractionation and

turnover of stable carbon isotopes in animal tissues: implications for δ13C analysis

of diet. Oecologia 57: 32-37.

Vander Zanden, M.J., Casselman, J.M., and Rasmussen, J.B. 1999. Stable isotope

evidence for the food web consequences of species invasions in lakes. Nature

401: 464−467.

Vander Zanden, M.J., and Vadeboncoeur, Y. 2002. Fishes as integrators of benthic and

pelagic food webs in lakes. Ecology 83: 2152-2161.

Vanderklift, M.A., and Ponsard, S. 2003. Sources of variation in consumer-diet δ15N

enrichment: a meta analysis. Oecologia 136: 169-182.

Vannote, R.L., Minshall, G.W., Cummins, K.W., Sedell, J.R., and Cushing, C.E. 1980.

The River Continuum Concept. Canadian Journal of Fisheries and Aquatic Sciences

37: 130-137.

Wallace, J.B., Eggert, S.L., Meyer, J.L., and Webster, J.R. 1997. Multiple trophic levels

of a forest stream linked to terrestrial litter inputs. Science 277: 102-104.

3.7 Tables and Captions

13 15 Table 3.1 Stable isotope ratios (δ Cadj and δ N, in ‰) and C:N for A. remigis females (F), males (M) and nymphs (N) collected over the course of the growing season in 2007 in four streams in New Brunswick, Canada. Each point represents a single pooled sample. Table continued on next page.

13 15 δ Cadj (‰) δ N (‰) C:N Site Date F M N F M N F M N Corbett Brook 7-May -26.7 -26.7 5.9 5.3 4.2 4.2 22-May -27.6 -27.1 4.5 5.3 4.3 4.2 11-Jun -27.3 -26.7 5.2 6.0 4.2 4.1 21-Jun -27.2 -26.8 5.2 5.3 4.2 4.2 4-Jul -26.7 -26.8 -26.6 5.7 6.0 3.7 3.9 4.0 3.7 17-Jul -26.3 4.6 3.6 2-Aug -25.7 -26.0 -25.8 4.7 4.7 4.7 4.4 4.2 4.1 14-Aug -26.6 -26.8 -26.0 5.2 4.9 3.2 4.1 4.1 3.7 28-Aug -26.1 -26.6 4.6 5.0 4.8 4.6 17-Sep -27.4 -27.1 -26.3 5.5 5.6 4.3 4.9 4.7 4.0 10-Oct -26.2 -26.8 4.7 5.3 5.5 5.0 Average (S.D.) -26.8 (0.6) -26.7 (0.3) -26.2 (0.3) 5.1 (0.5) 5.3 (0.4) 4.1 (0.7) 4.4 (0.5) 4.3 (0.3) 3.8 (0.2)

English Brook 7-May -26.6 -26.6 5.2 4.9 4.6 4.4 22-May -27.4 -26.5 4.8 4.6 4.3 4.2 11-Jun -26.2 -26.3 5.0 5.3 4.2 4.1 21-Jun -26.6 -26.1 4.9 5.8 4.0 4.2

4-Jul -25.8 -25.4 5.5 7.2 4.5 3.9 17-Jul -25.8 -25.9 -25.9 6.0 5.8 5.3 4.0 3.8 3.4 2-Aug -26.3 -26.6 -26.2 5.7 5.9 4.4 3.7 3.8 3.8 14-Aug -26.9 -26.0 -26.5 4.4 4.9 4.7 4.2 4.0 4.1 14-Sep -26.3 -26.5 5.3 4.8 5.4 5.4 10-Oct -25.8 -26.1 5.9 5.6 5.8 5.4 Average (S.D.) -26.4 (0.5) -26.2 (0.4) -26.2 (0.3) 5.3 (0.5) 5.5 (0.8) 4.8 (0.5) 4.5 (0.7) 4.3 (0.6) 3.8 (0.3)

McKenzie Brook 7-May -26.8 -26.2 5.2 5.1 4.5 4.1 22-May -26.7 -26.8 4.8 5.4 4.1 4.3 11-Jun -26.9 -26.1 5.2 5.5 4.2 4.2 21-Jun -26.0 -25.8 5.9 7.1 4.1 3.8 4-Jul -25.9 -26.5 7.2 5.3 4.0 4.1 17-Jul -26.1 -26.6 -25.9 6.8 6.3 5.8 4.0 3.8 4.0 2-Aug -25.7 -25.1 -25.8 5.7 6.1 5.4 4.2 3.7 4.3 14-Aug -25.9 -25.6 -25.8 5.6 6.2 5.7 4.3 4.0 4.1 30-Aug -26.2 -26.4 -26.2 6.1 5.2 4.1 5.3 4.3 3.9 17-Sep -26.4 -25.8 6.7 5.3 5.3 3.8 Average (S.D.) -26.2 (0.4) -26.1 (0.5) -25.9 (0.2) 5.8 (0.8) 5.9 (0.7) 5.3 (0.7) 4.3 (0.4) 4.2 (0.5) 4.0 (0.2)

Parks Brook 7-May -28.6 -28.4 6.2 6.3 4.3 4.2 22-May -29.1 -28.1 6.5 6.0 3.9 4.2 11-Jun -27.5 -27.5 6.7 6.6 4.2 4.1 21-Jun -28.9 -27.9 6.3 6.8 4.0 4.1 3-Jul -28.4 5.9 4.1

100

17-Jul -30.7 -30.7 -31.0 6.9 6.9 6.2 4.1 3.9 4.0 2-Aug -30.0 -31.9 -28.4 6.6 6.1 6.3 4.4 5.1 3.5 14-Aug -27.5 -28.1 7.0 7.0 5.2 5.2 30-Aug -27.9 -26.5 6.6 6.2 5.8 4.2 17-Sep -27.3 -26.6 6.3 7.2 5.7 5.1 Average (S.D.) -28.7 (1.2) -28.4 (1.7) -29.5 (1.9) 6.6 (0.3) 6.6 (0.4) 6.1 (0.2) 4.5 (0.6) 4.6 (0.7) 3.9 (0.3)

101

3.8 Figures and Captions

60 (-0.035t) 50 F weight = 49.3 - 41.6e r2 = 0.962 40 30 (-0.052t) 20 M weight = 35.9 - 27.9e 2

Wet weight (mg) Wet weight r = 0.961 10 0 0 1020304050607080 7 Experiment Day

5 C/N 4

0 1020304050607080 Experiment Day

Figure 3.1 Wet weight (A, in mg, mean of multiple individuals) and C/N (B, individuals) of Aquarius remigis nymphs (circles), adult females (squares) and adult males

(diamonds) vs. time in a laboratory diet-switch experiment. Animals were captured from a stream on day 0 and placed on a diet of crickets. Derived growth equations for female

(F, n = 14) and male (M, n = 13) striders are shown in (A).

102

-22 -23 -24 C

13 -25 δ 13 (-0.653t) -26 δ C = -24.23 - 2.71e r2 = 0.68 -27 half-life = 1.1 days -28 0 1020304050607080 Experiment Day

-22 -23 -24 adj

C -25 13 (-0.472t) 13 Cadj = -23.44 - 2.95e δ δ -26 r2 = 0.78 -27 half-life = 1.5 days -28 0 1020304050607080 Experiment Day

8 7

6 15 (-0.089t) δ N = 6.39 - 2.86e N 2 15 5

δ r = 0.93 4 half-life = 7.8 days 3 2 0 1020304050607080 Experiment Day

13 13 15 Figure 3.2 δ C (A, in ‰), δ Cadj (B, in ‰), and δ N (C, in ‰) in individual Aquarius remigis nymphs (circles), adult females (squares) and adult males (diamonds) vs. time in a laboratory diet-switch experiment (n = 52 for all three panels). Animals were captured

103

from a stream on day 0 and placed on a diet of crickets (dashed line indicates isotope ratio of the diet). Open symbols = tank 1 (n = 2/date), solid symbols = tank 2 (n =

2/date), solid triangle = mean value (± S.D.) for a sample collected from the source stream on day 51.

104

120 % Aquatic = 14.2*Order - 23.1 100 r2 = 0.33 80 p < 0.001 60 40 20 0 % AquaticCarbon in Diet -20 01234567 Stream Order

Figure 3.3 Percent aquatic carbon (mean values as estimated from mixing models using stable carbon isotopes) in the diet of Aquarius remigis (solid symbols) and Metrobates hesperius (open symbols) vs. stream order in New Brunswick, Canada streams. The best fit equation for A. remigis (n = 41) is included.

105

120 100 80 60 40 20 0 % AquaticCarbon in Diet -20 100May 150 June July 200 Aug 250Sept Oct 300 Date

Figure 3.4 Percent aquatic carbon (mean values as estimated from mixing models using

stable carbon isotopes) in the diet of Aquarius remigis (solid symbols) and Metrobates

hesperius (open symbols) vs. sampling date in Parks Brook, New Brunswick, Canada.

Squares = females, diamonds = males, circles = nymphs.

106

Chapter 4.0 Factors Affecting Water Strider (Hemiptera: Gerridae) Mercury

Concentrations in Lotic Systems

Timothy D. Jardine, Karen A. Kidd, Richard A. Cunjak, and Paul A. Arp

Submitted to Environmental Toxicology and Chemistry 2008

Abstract

Water striders (Hemiptera: Gerridae) have been considered as a potential sentinel for

mercury (Hg) contamination of freshwater ecosystems, yet little is known about factors

that control their Hg concentrations. Water striders were collected from a total of 80

streams and rivers in August and September in New Brunswick, Canada over a four year

period (2004-2007) to assess the influence of factors such as dietary habits, water

chemistry and proximity to point sources on Hg concentrations in this organism. Spatial

patterns of high and low Hg concentrations were consistent with previous observations

and expectations based on point source emissions, with higher than average

concentrations observed in the southwest and Grand Lake regions of the province. Total

Hg concentrations in this latter region were elevated (total Hg > 0.40 ug g-1) in the lichen

Old Man’s Beard (Usnea spp.) within 10 km of a coal-fired power plant that is a

significant source of Hg (~100 kg annually), while water strider Hg concentrations

peaked (total Hg > 0.50 ug g-1) between 10 and 50 km from the plant. Across all streams, pH and total organic carbon of water were relatively weak predictors of water strider Hg

concentrations. Female water striders that were larger in body size than males had

significantly lower Hg concentrations within sites, suggestive of growth dilution. Percent

107

aquatic carbon (using δ13C) in the diet of striders ranged from 0 to 100% and there was

no relationship between % aquatic carbon in the diet and Hg concentrations in striders.

For those striders feeding solely on terrestrial carbon, Hg concentrations were higher in

animals occupying a higher trophic level (using δ15N). Hg concentrations in striders collected monthly over the growing season in two years were highly variable, suggesting that concentrations are likely to reflect short-term changes in Hg availability. These measurements highlight the importance of considering both deposition and post-

deposition processes in assessing Hg bioaccumulation in this species. They also suggest

water striders may be more appropriate as a terrestrial than aquatic Hg sentinel, underscoring the importance of understanding the origin of food for organisms used in

contaminant studies.

4.1 Introduction

Mercury (Hg) contamination of freshwater ecosystems remains a major problem

in industrialized nations, with deposition from natural and anthropogenic sources and

subsequent methylation leading to high Hg concentrations in fishes (Boening 2000).

Human and wildlife health concerns surrounding consumption of Hg-contaminated fish requires considerable research and monitoring efforts by government agencies

(Cunningham et al. 1994) and losses of Hg-contaminated products (mainly fish and

marine mammals) are estimated at billions of dollars globally (Hylander and Goodsite

2006).

Environmental sentinels, including lichens (Garty 2001), invertebrates (de Freitas

Rebelo et al. 2003), fishes (Munkittrick and Dixon 1989) and birds (Fox et al. 2002), hold

108

great promise in providing efficient and ecologically relevant information on the regional and global distribution of contaminants (Beeby 2001, Nummelin et al. 2007). Ideal characteristics of sentinel species include wide distribution, limited home range, a well- known life history, moderate to high abundance, and simple taxonomic identification

(Beeby 2001). The predaceous water strider (Hemiptera: Gerridae) meets many of these criteria. One species, Aquarius remigis Say, is common and abundant on the surface film of streams and rivers across North America (Andersen 1990), has a home range that is restricted to ~100 m (Fairbairn 1986), and has Hg concentrations that are correlated with those in small brook trout, an important recreational fish species (Jardine et al. 2005).

Despite an improved understanding of the Hg cycle over the past 40 years (Morel et al. 1998), several fundamental questions remain unanswered about the effect and extent of point sources on the surrounding environment, and the relative importance of atmospheric deposition compared with other abiotic and biotic processes such as water chemistry and food web characteristics. While mathematical models predict that reduced emissions from point sources such as coal-fired power plants will result in reductions in animal tissue concentrations (Evers et al. 2007) and certain studies have shown such concentration declines after emissions reductions (Hrabik and Watras 2002, Evers et al.

2007), these decreases may not always be achieved due to complexities associated with water chemistry and biology (Wong et al. 1997, Watras et al. 1998). Given the logistic and financial challenges surrounding the collection and analysis of fish samples for Hg on a broad geographic scale (Kamman et al. 2005), and quality control issues associated with bringing together fish Hg data analyzed in different laboratories, water striders are envisioned as a rapid means of assessing spatial patterns in Hg bioaccumulation in lotic

109

food webs (Jardine et al. 2005). Specifically, herein water striders were sampled in order to identify potential Hg “Hotspots” and “Areas of Concern”, as has been reported previously for Northeastern North America (Evers et al. 2007).

This study examined spatial variability and potential factors that influence Hg concentrations in water striders. First, preliminary sampling was conducted on a broad geographic scale in New Brunswick, Canada to assess the variation in strider Hg concentrations across the landscape. A study was then designed to assess spatial patterns in Hg deposition relative to a coal-fired power plant in New Brunswick, and determine if water strider Hg concentrations reflect local deposition. Lichen (Usnea spp., Old Man’s

Beard) were used as a 2nd sentinel species since they are indicators of heavy metal pollution via atmospheric deposition (Garty 2001). Based on principles outlined in

Schroeder and Munthe (1998), it was predicted that Hg concentrations would peak between 10 and 50 km from the power plant with the deposition of divalent Hg and decrease thereafter. A variety of other factors was examined as possible determinants of water strider Hg concentrations. For example, increased acidity (Gilmour et al. 1992) and organic matter content (Harding et al. 2006) of water, changes in growth and activity rates (Karimi et al. 2007, Trudel and Rasmussen 2006), and differences in feeding ecology (Kidd et al. 2003) can all modulate Hg concentrations in aquatic biota. Water chemistry characteristics such as pH, organic matter content and waterborne Hg were expected to have significant effects on Hg concentrations in water striders, while differences in growth patterns among water strider species and sexes within sites were also examined to explain possible differences in Hg concentrations across sites. Hg data from all sites were also compared with previously reported (Chapter 3) stable isotope

110

ratios of carbon (13C/12C or δ13C) and nitrogen (15N/14N or δ15N) as indicators of food

source pathway (aquatic vs. terrestrial) and trophic level, respectively, to assess the

effects of feeding ecology on Hg concentrations in striders. Finally, seasonal and inter-

annual changes in Hg concentrations were measured at four index sites sampled bi-

weekly during the growing season over two years. All of these measurements were done

to assess the utility of water striders as environmental Hg sentinels, and determine the

relative importance of atmospheric deposition and in-stream processes in affecting Hg

concentrations in this organism.

4.2 Methods

Water striders were collected with hand nets over a four year period in 2004-2007

from a total of 80 streams in New Brunswick. In 2004 water striders were collected from

42 randomly selected streams (Fig. 4.1a) that represented the eight recreational fishing

areas designated by the provincial authority (New Brunswick Department of Natural

Resources) and the major drainage basins in the province; these data were used to

establish regional patterns of Hg concentrations. These sites were generally forested 1st

to 6th order streams and rivers with cobble/gravel bottoms (Appendix 1). Although four

species of water strider were present at some sites (Aquarius remigis, Metrobates hesperius, Limnoporus sp., and Gerris comatus), adult A. remigis was most common and collections focused primarily on this species. M. hesperius was also collected at nine of the sites in 2004, but it only occurred in sympatry with A. remigis at one site.

In 2005, A. remigis was collected from 13 streams, with preliminary sampling directed at locations near two potential point sources: coal-fired power plants operated by

111

NB Power in Belledune (northern NB) and Grand Lake (south-central NB) (Fig. 4.1a).

The Grand Lake plant, burning local high S coal (S content = 6.6%, Zhao et al. 1997) at about 230,000 to 430,000 metric tons per year, is a source of low-grade sulphur dioxide

(Meng et al. 1995) and also emits approximately 100 kg of Hg per annum, roughly 4% of all Hg released by anthropogenic sources in Canada (Environment Canada 2005). The

Belledune plant releases ~13 kg of Hg per annum (Environment Canada 2005) but the presence of two other local sources (a metal smelter emitting ~30 kg Hg per annum and a chlor-alkali plant emitting ~32 kg Hg per annum, Sensen and Richardson 2002,

Environment Canada 2005) add further amounts of Hg to the local environment.

In 2006 a bullseye design (Green 2005) was used to select sampling sites and map

Hg concentrations in water striders and Usnea spp., with the Grand Lake generating station at the centre of the bullseye (Fig. 4.1b). Repeat sampling of the sites within the bullseye was done in 2007, due to an operational shutdown at the station in 2006 during the months of July and August (A. Bielecki, New Brunswick Power Corp., pers. comm.).

A minimum of one site was chosen within each section delineated by six radii (10, 20,

30, 50, 100, >100 km) and eight compass directions, yielding a total of 60 sites (Figure

4.1b). Water striders were collected in August/September from these sites, with A. remigis again being the main target species (collected at 51 sites); M. hesperius was also collected when present (9 sites). Old Man’s Beard was randomly sampled from 2-5 trees per site at the same locations sampled for water striders (Fig. 4.1b), and pooled into a single composite for analysis, for a total sample size of 60.

In all four years, water quality samples were collected during baseflow conditions

(August/September, n=1/site). In 2004 water samples were analyzed for total Hg and

112

total organic carbon with a Tekran Model 2600 (US EPA Method 1631) and a Technicon

Traacs 800 Auto Analyzer, respectively. In 2005, water samples were analyzed for total

Hg at the Environment Canada Laboratory in Moncton, NB, by flameless atomic absorption spectrometry after oxidation to inorganic mercury by sulphuric acid, dichromate and UV photo-oxidation and reduction with stannous sulphate in hydroxylamine sulphate - sodium chloride solution. The detection limit was 0.02 ug/L.

From 2005-2007, water samples were analysed for pH and total organic carbon at the

New Brunswick Department of Environment Analytical Laboratory (Fredericton, NB).

Due to logistic constraints associated with the collection of large volumes of water at remote sites, in 2006 and 2007 water samples were not analyzed for Hg. Also, because the pH, TOC, and Hg of water samples was not being compared among years, no trials were performed to compare data generated by the different techniques.

In 2006 and 2007, water striders from four index sites (Corbett Brook, N 45.92 W

66.64, English Brook N 46.43 W 66.60, McKenzie Brook N 46.22 W 66.53, and Parks

Brook N 45.46 W 66.35) were sampled bi-weekly from May to October to assess seasonal changes in Hg concentrations. A. remigis adults and nymphs were collected at all four sites; M. hesperius was only present at Parks Brook and was not included in any temporal analyses.

For the strider samples collected in both 2004 and 2005, two to three composite samples of two individuals (sexes and wet weights not determined) per composite were analyzed for total Hg from each site. In 2006 and 2007, male and female striders were analyzed separately; males are readily discernable from females by inspection of genital morphology (Fairbairn et al. 2003). For each site, individuals were weighed to obtain

113

wet weights and 2-7 individuals were pooled for each sex. This yielded a mean wet

weight and a single composite Hg concentration for each sex at each site on a given date.

In the laboratory, water strider samples were freeze dried for a minimum of 48

hours and homogenized with a mortar and pestle. Lichen samples were cut with stainless

steel scissors into 1 cm sections and freeze-dried prior to analysis. All instruments were

cleaned with 10% HCl between samples. All strider and lichen samples (2004-2007)

were weighed to ~10 mg and analyzed for total Hg using a Direct Mercury Analyzer

(DMA 80, Milestone Microwave Laboratory Systems, Shelton, CT). Data are reported as

ug g-1 dry weight. Samples were run with certified reference materials DORM-2 (dogfish

muscle, National Research Council, Ottawa, ON, Hg = 4.64 ug g-1) and IAEA 336

(lichen, International Atomic Energy Agency, Vienna, Austria, Hg = 0.20 ug g-1) for water striders and Old Man’s Beard, respectively. Recoveries of DORM-2 and IAEA

336 were 93.3 ± 3.6% S.D. (n = 60) and 78.2 ± 1.6% S.D. (n = 17), respectively. The low value for IAEA-332 is likely due to the aliquot received, as the same batch analyzed at a 2nd lab yielded similar results (76.5 ± 3.7% S.D., n = 33). These values are at the

lower end of the certified range for this standard. Sample repeats yielded average

standard deviations of 0.02 ug g-1 both within (n = 24) and across (n = 25) analytical runs.

To determine if the power plant was causing increased deposition of sulphur (Meng et al.

1995) and whether S and Hg concentrations exhibited similar deposition patterns, Old

Man’s Beard samples from 2006 were also analyzed for % S using a LECO CNS 2000

elemental analyzer (LECO Instruments Ltd., Mississauga, ON).

Mercury concentrations in water striders were compared to their dietary habits

using previously-reported % aquatic carbon data (based on δ13C) and δ15N as described in

114

Chapter 3. For analysis of stable isotopes, approximately 0.2 mg of each freeze-dried,

ground sample was weighed into tin cups. Samples were analyzed with a NC2500 elemental analyzer connected to a Finnigan Delta XP mass spectrometer. Isotope data

are expressed using delta notation in per mil (‰) according to: (Rsample/Rstandard –

1)*1000, where R is the raw ratio of heavy to light isotope (e.g. 13C/12C) and standards are Peedee belemnite carbonate and atmospheric nitrogen for carbon and nitrogen, respectively. Accuracy and precision were monitored with commercially available standards as described in Chapter 3.

A subset of water strider samples (A. remigis n = 26, M. hesperius n = 6) was analyzed for methyl Hg at the University of Ottawa. These samples were selected from

14 sites sampled in 2005 and 2006 and approximated the range of total Hg concentrations observed (0.05 to 0.71 ug g-1). Methyl Hg extractions were done using methods outlined

in Al-Reasi et al. (2007). Resultant solutions were analyzed by GC-Mass Spectrometry

on a HP 6890 series with HP injector series 7683 following Cai et al. (1997). Recovery

of a certified reference material (DORM-2) averaged 98 ± 12% S.D. (range = 83 to

118%, n = 13).

Data were analyzed using NCSS (Kaysville, UT) and SYSTAT (version 9, SPSS,

Inc., Chicago, IL) software. All Hg data were log-transformed prior to analysis to reduce

heteroscedasticity. Based on inspection of a probability plot, % sulphur data were judged

to be normally distributed and hence were not transformed. Concentrations of methyl

versus total Hg were compared between sexes and between species of water striders

using analysis of covariance (ANCOVA) with total Hg as the covariate. To assess

general patterns of strider Hg concentrations from sampling done in 2004 and 2005,

115

analysis of variance (ANOVA) was used with site nested within recreational fishing area

(2004) and site nested within region (2005). For Old Man’s Beard data, linear

regressions were used with distance as the independent variable and Hg concentrations

(2006 and 2007) or % sulphur (2006 only) as the dependent variable. ANCOVA, with direction (NW, NE, SE, SW) as a factor and distance from the power plant as a co- variate, was used to test the effect of compass direction separately in 2006 & 2007 for Hg in Old Man’s Beard and in male and female water striders. It was also used to test the effect of year (2006 & 2007) on Old Man’s Beard Hg concentrations with distance from the power plant as the co-variate. A similar analysis with water strider males and females both yielded significant year x distance interactions. However, ANCOVA, with sex as a factor and distance as a co-variate, did not yield interactions and was used to test the effect of distance from the power plant on water strider Hg concentrations separately in

2006 and 2007. A multiple regression model was also used with pH and distance from the power plant as variables because pH and distance were highly correlated in both years

(Appendix 2). Two sites were excluded in 2007 as outliers in Hg vs. distance plots, identified by R[student] > 2. Linear regressions were used to test the effect of different water chemistry variables and % aquatic carbon on strider Hg concentrations (focusing on females in 2006 and 2007). For those sites with no contribution of aquatic carbon to strider diet (% aquatic carbon = 0%), Hg concentrations in striders were compared to their δ15N as an indicator of trophic level. For this analysis it was assumed there was no

variation in baseline δ15N across sites since terrestrial organic matter shows little

variation (e.g. alder, the most common riparian tree species at the sites: δ15N = -1.0 ±

0.5‰ S.D., range = -2.6‰ to 1.4‰, n = 92, T.D.J., unpubl. data). Because data were

116

from pooled samples, paired t-tests were used to compare Hg concentrations and body weights between sexes within species, to compare Hg concentrations between A. remigis and M. hesperius at sites where they co-existed, and also to compare Hg concentrations in striders between years as an alternative to the ANCOVA that yielded interactions. Sites were used to pair sexes, species and years in these three analyses.

For the data collected at the four index sites over the growing season, results were grouped into two generations (as in Chapter 3) - the first that returned to stream surfaces after overwintering (“post-winter”) and the new generation that hatched in early summer

(“pre-winter”). Adults sampled in the period after ice-out in the spring up to and including the first day when a new generation of nymphs was present (typically early

July) were considered as the post-winter sample. Adults sampled for the remainder of the growing season (July – October) were made up of a new generation and considered pre- winter. This latter period also corresponded roughly to the time period when the one- time spatial sampling (bullseye) was conducted. The timing of the arrival of the new generation, and hence the delineation of the two samples, varied slightly from one stream to another (see Fig. 4.7 for details). Originally, the intention was to examine whether Hg concentrations varied significantly over time and between sexes at the four streams that were regularly sampled by comparing adult water strider Hg concentrations across sampling periods using general linear model analysis of variance (GLM-ANOVA). The

GLM-ANOVA had three factors – site (random factor: Corbett Brook, English Brook,

McKenzie Brook, and Parks Brook), sex (fixed factor: male and female), and generation

(fixed factor: post- and pre-winter) separately for the two years of data (2006 & 2007) with a Bonferroni-adjusted probability of 0.025. However, because several interactions

117

were significant (p < 0.025) in this model, including site x generation and sex x generation, data were further separated into four groups as two generations in each of

2006 and 2007. Within each of these groups, a GLM-ANOVA was run with two factors

– sex (fixed) and site (random). Bonferroni adjusted probabilities were used with α =

0.05/4 = 0.012.

To compare Hg concentrations in adults and nymphs, average values for males and females were compared to those of nymphs for those times that they co-occurred

(mainly summer) with two factors (stage – fixed, and site - random) in a GLM-ANOVA with α = 0.05.

4.3 Results

Samples collected in 2004 and 2005 revelaed regional patterns in water strider Hg concentrations in New Brunswick. In 2004, total Hg concentrations in water striders ranged from 0.08 to 0.52 ug g-1 across the 31 sites. Concentrations were typically more elevated at sites in the southwest portion of the province while lowest concentrations occurred in the northern half of the province (Table 4.1, Fig. 4.1). There were no significant differences among recreational fishing areas (p = 0.064), but strong differences among sites (p < 0.0001) in this year. In 2005, total Hg concentrations in striders were significantly different among sites (p < 0.001) and had a similar range (0.06 to 0.52 ug g-1) across sites as those collected in 2004. Concentrations were significantly elevated in the Grand Lake region compared to the Belledune region (p=0.047), despite a site immediately adjacent to the Grand Lake power plant having striders with the lowest concentration recorded (0.06 ug g-1) (Table 4.1).

118

Most of the Hg in water strider tissues was in the form of methyl Hg (% methyl

Hg = 87 ± 15% S.D., range = 59 to 132%, n = 32, both species combined; values >100%

stem from analytical error associated with methyl and total Hg determination), and there

were no differences in % methyl Hg between male and female A. remigis (p = 0.972) or

between species (p = 0.382). A best fit equation relating methyl Hg to total Hg was:

Methyl Hg = 0.865*Total Hg + 0.0025 (r2 = 0.96).

Mercury concentrations in Old Man’s Beard ranged from 0.06 ug g-1 to 0.52

ug g-1 and declined significantly with distance from the power plant in both 2006 and

2007 (p < 0.001, Fig. 4.2a). Sulphur concentrations in 2006 also declined with distance

from the plant (p = 0.007, Fig. 4.2b), although the effect was not as strong as that

observed for Hg (Hg vs. distance r2 = 0.41 in 2006, 0.29 in 2007; %S vs. distance r2 =

0.13, Fig. 4.2). There were no differences in Hg concentrations between years for Old

Man’s Beard (distance: p < 0.001; year: p = 0.864). In 2006, there were significant differences among directions in Hg concentrations (p = 0.024), with sites to the northeast of the power plant having higher concentrations than sites to the southwest. In 2007, there were no significant differences among directions (p = 0.076).

In 2006 while the power plant was not operational, Hg concentrations in water striders showed no linear relationship with distance from the generating station (p =

0.772) for either females (p = 0.902, n = 52) or males (p = 0.592, n = 51) (Fig. 4.3a). In addition, there was no correlation between Hg concentrations in Old Man’s Beard and concentrations in male (r < 0.01 , p > 0.05) or female (r = -0.16, p > 0.05) water striders across sites for this year. When comparing data from 2006 to 2007, Hg concentrations significantly increased in males (p = 0.039) but not females (p = 0.093). In contrast with

119

2006, Hg concentrations in striders collected in 2007 significantly declined with distance

from the plant (ANCOVA, p < 0.001) in both females (p = 0.025, n = 49) and males (p =

0.002, n = 46) (Fig. 4.3b). There was no effect of compass direction from the power

plant on Hg concentrations in males or females in either 2006 or 2007 (p > 0.05). When

analyzed in a multiple regression with pH and distance as variables, trends were generally

similar to those observed with distance analyzed alone. Stream-water pH had a weak but

significant effect on female strider Hg concentrations in 2006 (r2 = 0.10, p = 0.026) while

distance had no effect (r2 = 0.02, p = 0.310) and males showed no effect of pH (r2 = 0.06, p = 0.097) or distance (r2 = 0.003, p = 0.682). In 2007, female Hg concentrations showed

no effect of pH (r2 = 0.05, p = 0.122) or distance (r2 = 0.04, p = 0.137) but males showed

a significant effect of distance (r2 = 0.10, p = 0.025) and not pH (r2 = 0.04, p = 0.138).

When two outliers were included in the analysis, relationships between distance and Hg

concentrations and pH and Hg concentrations were non-significant (p > 0.05) for males

and females in both 2006 and 2007.

Water quality variables were inconsistent predictors of Hg concentrations in water

striders over the four years of study, accounting for a maximum of 64% of the variation

when analyzed independently (Table 4.2). In 2004, significant relationships were

observed between Hg in A. remigis and pH (r2 = 0.38, p < 0.001), TOC (r2 = 0.14, p =

0.030), and conductivity (r2 = 0.22, p = 0.003). There was no relationship between Hg in

striders and total Hg in water in 2004 for either strider species, but in 2005 a significant

positive relationship between these two variables was observed for A. remigis (r2 = 0.64, p = 0.008). TOC (r2 = 0.41, p = 0.008) and conductivity (r2 = 0.51, p = 0.003) were also

significant predictors of Hg in A. remigis in 2005 (Table 4.2). In 2006 and 2007, none of

120

the water quality variables were significant predictors of Hg concentrations in A. remigis or M. hesperius (Table 4.2).

In both 2006 and 2007 there were between-sex differences in body size and Hg concentrations for both species of water striders. Female water striders were consistently larger than males (Figs 4.4a and 4.5a). The difference in body weights between sexes was less pronounced in A. remigis (2006: mean % difference = 23.0 ± 10.8% S.D., n =

54, p < 0.001; 2007: mean % diff. = 21.2 ± 14.8% S.D., n = 47, p < 0.001; Fig. 4.4a) than

M. hesperius (2006: mean % difference = 61.9 ± 10.3% S.D., n = 9, p < 0.001; 2007: mean % diff. = 65.4 ± 6.7% S.D., n = 7, p < 0.001; Fig. 4.5a). Male striders had higher

Hg concentrations than females for both A. remigis (2006: mean % difference = 13.4 ±

31.9% S.D., n = 50, p = 0.001; 2007: mean % diff. = 3.8 ± 7.6%, n = 47, p < 0.001; Fig.

4.4b) and M. hesperius (2006: mean % diff. = 53.7 ± 36.3% S.D., n = 9, p < 0.001; 2007: mean % difference = 58.1 ± 8.6% S.D., n = 7, p < 0.001; Fig. 4.5b). There was no correlation between % difference in weight and % difference in Hg concentration between sexes within sites for A. remigis (2006: r2 = 0.02, p = 0.750; 2007: r2 = 0.02, p =

0.400) or M. hesperius (2006: r2 = 0.42, p = 0.058; 2007: r2 = 0.06, p = 0.648). M. hesperius had significantly higher Hg concentrations than A. remigis (n = 6 sites, p =

0.003, data not shown) where the two species co-existed.

There was no relationship between % aquatic carbon in the diet and Hg concentrations in A. remigis water striders (p > 0.05, Fig. 4.6a), however these analyses were limited given the large number of sites (n = 22 of 41, Chapter 3) with striders having 0% aquatic carbon in the diet. For those striders that had aquatic carbon in their diets (2 to 100 %), their Hg concentrations fell within the range of those animals that

121

relied solely on terrestrial carbon with one exception. Striders from a single site (Clark

Brook, N 46.06 W 65.54) had atypically high Hg (average Hg = 2.0 ug g-1) and were

feeding mainly on aquatic food sources (average % aquatic carbon = 79%). For those

striders with 0% aquatic carbon in the diet (i.e. entirely terrestrial feeders), δ15N as a measure of trophic level explained significant variation in Hg concentrations, with high

δ15N (high trophic level) associated with high Hg (p = 0.001, r = 0.60, Fig. 4.6b). By

including distance from the power plant as a 2nd variable in a stepwise regression, over

half of the variation was accounted for (r2 = 0.53), and both δ15N (p = 0.001) and distance

(p = 0.012) were significant.

Seasonal and inter-annual variation in A. remigis Hg concentrations at the four

index sites was high (Fig. 4.7), but patterns in three of the four streams (exception

Corbett Brook) were similar, with generally lower concentrations in the late summer/fall

compared to spring. At Corbett Brook, the 2nd generation of 2006 (adults from July to

October) had high Hg concentrations (0.40 to 0.60 ug g-1) that remained high into the

spring of 2007 after overwintering; in contrast, the 2nd generation collected at this site in

2007 then had lower concentrations (0.10 to 0.30 ug g-1, Fig. 4.7). During the pre-winter

period when one-time sampling at 60 sites was conducted, the other three index sites had

consistent concentrations between years. At all sites, concentrations of Hg in males and

females diverged in the spring, with male concentrations increasing relative to females

(Fig. 4.7). Concentrations of Hg in females generally remained low (0.10 to 0.30 ug g-1) throughout the growing season. Overall, there were several interactions (p < 0.025) between sexes, generations and sites. When analyzed separately by generation and year, post-winter males had significantly higher Hg than post-winter females in both 2006 (p =

122

0.002) and 2007 (p = 0.006). Hg concentrations among sites were not significantly

different during the post-winter sample in 2006 (p = 0.275) but they were different in

2007 (p = 0.004). For pre-winter samples, Hg concentrations among sexes were not

significantly different in either year (p > 0.100) but among sites were different during this

time period in both years (2006: p < 0.001; 2007: p = 0.010). In 2006, nymphs had

similar concentrations to adults (p = 0.086); in contrast, nymphs collected in 2007 had

significantly lower Hg concentrations than those of adults (p = 0.013, Fig. 4.7).

4.4 Discussion

This study examined the physical, chemical and biological factors affecting Hg

concentrations in water striders collected across the province of New Brunswick (NB).

Variation in water strider Hg concentrations was related to water chemistry, distance

from the coal fired generating station, sex, body size, and trophic level, highlighting the

importance of both abiotic and biotic factors in determining Hg concentrations in this

organism. Water striders revealed regional patterns of Hg concentrations in New

Brunswick that may be related to atmospheric Hg deposition and landscape characteristics, with striders from the northern part of the province having lowest Hg concentrations and those from the southern part having highest concentrations. These

organisms feed mainly on terrestrial carbon (>50%; Chapter 3); this likely explains the weaker relationship between water chemistry parameters and their Hg concentrations,

and suggests that striders may be more useful as indicators of Hg availability in the

terrestrial rather than aquatic environment.

123

Spatial studies with sentinel species are useful for identifying areas of high and low contaminant concentrations (Beeby 2001). In this study, Hg concentrations in striders collected in 2004 generally decreased in a southwest to northeast direction (Fig.

4.1 and Table 4.1); this pattern is consistent with studies on other organisms in the region

(Evers et al. 2007) and earlier work with striders there (Jardine et al. 2005). Evers et al.

(2007) found yellow perch (Perca flavescens) and loons (Gavia immer) from lakes in the

Lepreau Region (southwest NB) had elevated Hg concentrations. In this study, Hg analyses of water striders and Old Man’s Beard in 2006 and 2007 also showed elevated concentrations (>0.40 ug g-1) in the Grand Lake region in south-central NB (Fig 4.2, 4.3).

Areas characterized by higher concentrations of Hg in either water striders or Usnea spp

may potentially contain fish with high Hg concentrations, and therefore could be targeted

as locations for more detailed food web sampling for Hg.

Concentrations of Hg in the sentinel species varied considerably among sites that

were at similar distances from the coal fired generating station and this may be due to

spatial variability in the atmospheric deposition of this pollutant. Prior monitoring and

modelling efforts for the Grand Lake power plant revealed that S deposition was affected

by: 1) the prevailing wind direction, which is generally directed towards the northeast; 2)

the topography, with greater S deposition occurring on the ridges than valleys located

along the northeast direction; 3) the plume height, as overall S deposition rates can be

expected to decrease with increasing plume height; and 4) variations in atmospheric

stability, with highest S deposition patterns occurring during unstable and neutral

conditions (Bourque and Arp, 1994, Meng et al. 1995). Deposition of Hg in this area

124

could also be affected by these processes and explain some of the among-site differences

for both striders and Usnea spp. at comparable distances from the generating station.

We found increased concentrations of Hg in striders collected near the plant (~10-

50 km away) in 2007 when compared to 2006, possibly due to the unanticipated

shutdown that occurred during the first sampling period. Given their strong linkages to

terrestrial carbon (Chapter 3), the main source of Hg for water striders would be from Hg

in terrestrial insects, which could respond relatively quickly to changes in deposition rates via transfer from terrestrial vegetation (Hintelmann et al. 2002). Links have been shown to exist between precipitation amounts, Hg deposition from precipitation, and Hg concentrations in biota. For example, wet deposition of Hg (in ug/m3) is correlated with

annual precipitation in both Japan and North America, with wetter areas receiving higher

fluxes of Hg (Sakata and Marumoto 2005). Hammerschmidt and Fitzgerald (2005, 2006)

found a strong correlation between annual wet Hg deposition and methyl Hg in

mosquitoes and fish. This suggests that on relatively large spatial and temporal scales,

Hg deposition may predict Hg in biota (Miller et al. 2005). Determining the relationship

between Hg deposition and Hg in water striders will require sampling across a much

broader range of Hg deposition, such as that observed across North America

(Hammerschmidt and Fitzgerald 2005, 2006), than that represented in this study.

In 2006 when striders and lichen were sampled during a shutdown of the

generating station (A. Bielecki, NB Power corporation, pers. Comm.), only lichen

showed decreasing Hg concentrations with increasing distance from the plant (Fig 4.2).

In contrast, in 2007 when the plant was operational both lichen and striders showed

higher Hg concentrations at sites closer to the plant. Differences in lifespan of these

125

sentinel species and responsiveness to changes in atmospheric deposition may explain the among-year variability. Consistently high Hg concentrations in Usnea spp. nearest the source may simply reflect a longer lifespan (Keon 2001, Sancho and Pintado 2004) and

Hg accumulation from previous years or decades when the power plant was operating at a greater capacity, burning coal with a higher concentration of Hg, or not yet using emission control technology. Striders, meanwhile, have a one-year lifespan (Fairbairn

1985) and exhibit rapid turnover of their body tissues (Chapter 3); hence they are more likely to reflect recent exposure to Hg.

While Hg concentrations in striders and lichens were higher closer to the coal- fired generating station at Grand Lake (Fig. 4.2, 4.3), maximum concentrations of Hg occurred at different distances for these two sentinels. The zone of influence indicated by lichens (~0-10 km from the power plant) is comparable to that found for a chlor-alkali plant in New Brunswick (Sensen and Richardson 2002) and for ground level measurements of gaseous Hg around a chlor-alkali plant in Sweden (Wangberg et al.

2005), but strider Hg concentrations peaked at distances 10-50 km from the generating station and were poorly correlated with Hg concentrations in Usnea spp. This suggests some differences in Hg exposure for the two sentinels, even though atmospheric deposition is expected to exert some control over Hg concentrations in both Usnea spp.

(via direct uptake) and striders (indirectly via vegetation and riparian insects given the importance of terrestrial carbon in their diets, Chapter 3). Although biogeochemical cycling of Hg is not well understood and the spatial differences between these two sentinel species cannot be explained at present, the variability in Hg concentrations may be due to the fact that lichen take up Hg directly either through gaseous or particulate-

126

bound forms whereas Hg in striders would be affected by various chemical and biological processes after it is deposited onto the terrestrial or aquatic landscape and taken up into its prey. It is also possible that their concentrations reflect the deposition of different forms of Hg; lichen concentrations may be more reflective of particulate-bound Hg deposited closer to the generating station whereas strider concentrations may reflect higher deposition of gaseous forms of Hg2+ at distances further removed from the power plant (Schroeder and Munthe 1998). In the Grand Lake Region, it is possible that deposition of Hg bound on particulates occurs closer to the power plant than the deposition of gaseous Hg2+. Old Man’s Beard situated to the northeast of the power plant

had higher concentrations than those to the southwest in 2006, consistent with the prevailing wind direction for the area and previously measured patterns of SO2 deposition (Bourque and Arp 1994, Meng et al. 1995). This suggests that dry deposition was an important source of Hg to lichens because most precipitation (i.e. storm events) originates from the opposite direction, the northeast (Bourque 1992).

Results from this study concurred with others that have found lower than expected

Hg concentrations in aquatic consumers at sites immediately adjacent to emission sources

(Anderson and Smith 1977, Pinkney et al. 1997, Lipfert et al. 2005). Another possible

explanation for the spatial differences between strider and lichen Hg concentrations

described above may be due to decreased bioavailability of Hg to striders living closer to

the power plant. This decreased bioavailability may be due to an inhibition of

methylation of Hg or to the concurrent deposition of selenium (Belzile et al. 2005). For

example, lakes near Sudbury, Ontario metal smelters with high Se concentrations in

water have biota with lower total and methyl Hg concentrations than lakes far away from

127

the smelters with low Se concentrations (Belzile et al. 2005). Selenium can interfere with

Hg binding sites in proteins and thus limit Hg assimilation, as well as participating in the demethylation of methyl Hg (Iwata et al. 1982, Cuvin-Aralar and Furness 1991). While previous studies have not found unusually high Se concentrations in New Brunswick wildlife (Burgess et al. 2005, Braune and Malone 2006), these surveys were not conducted in the Grand Lake region. It is known, however, that Se co-accumulates with

S in sulfide-carrying coal beds such as those of the Grand Lake area (Yudovich and

Ketris 2006). The local burning of this coal with a high S content of 6.6% (Zhao et al.

1997) would therefore add Se as a logical associate to the local S and Hg emission and deposition patterns, but at concentrations 2000 times lower than that of the S emissions

(Interagency Monitoring of Protected Visual Environments 1994). Whether this deposition is sufficient to affect Hg uptake by striders remains to be resolved.

Water chemistry, particularly acidity and organic matter content, is typically a determinant of Hg concentrations in aquatic organisms (Suns and Hitchin 1990, Mason et al. 1996, Watras et al. 1998). Recent work on blackflies, which reside low on the food chain as primary consumers, showed strong relationships between Hg concentrations and pH and dissolved organic carbon (DOC) (Harding et al. 2006), likely because low pH and high DOC may increase the availability of Hg to lower trophic levels (Mason et al. 1996).

In the current study striders had Hg concentrations that were not consistently correlated with pH and TOC of the streams across sampling years; however, among-site variability in water chemistry may be less important for species such as water striders that derive the majority of their biomass from the terrestrial environment (Chapter 3) and are thus partly disconnected from processes occurring in the aquatic environment.

128

While relationships between Hg concentrations in water striders and stream water chemistry were inconsistent, there were consistent and significant differences within sites between sexes and species; within sites females of both species had larger body sizes and lower Hg concentrations than males. There is no evidence for differences in age between male and female striders reaching adulthood in late summer (Galbraith and Fernando

1977), meaning differences in size are most likely due to differences in growth (Fairbairn

1997) and differences in Hg between sexes suggest growth dilution of Hg by females

(Hammar et al. 1993, Simoneau et al. 2005, Karimi et al. 2007). Furthermore, M. hesperius attain smaller maximum body sizes than A. remigis (Fig. 4.4a, Fig. 4.5a) yet have higher Hg concentrations, suggesting a link between growth and Hg concentrations across species. Hg concentrations in fishes can be affected by differences in feeding rates, food conversion efficiency and growth (Trudel and Rasmussen 2006). Since A.

remigis populations in NB have a single generation each year (TDJ, unpubl. data,

Blackenhorn and Fairbairn 1995) and A. remigis mean adult body sizes differed between

sites by 72% in males and 66% in females, among-site differences in body size likely

reflect differences in growth and food availability. These growth differences among sites

could therefore confound assessment of spatial patterns in Hg concentrations and

contribute to some of the variability observed here.

The sex difference in growth and Hg for A. remigis is not due to source of food or

trophic level (Chapter 3), but could be due to differences in feeding rate, activity level or

loss during egg deposition. There are major differences between sexes in the spring in

activity levels, where males aggressively seek out female partners for copulation (Sih et

al. 2002, A. Sih, University of California, Davis, CA, pers. comm.). Increased activity

129

relative to food consumption can increase Hg concentrations (Trudel and Rasmussen

2006), but body size and Hg differences between sexes are also evident during the fall.

While it is also possible that females may lose part of their Hg burden via egg production,

nymph Hg concentrations were similar to breeding females at all four sites suggesting no

net loss of Hg through this pathway. During the late summer (pre-winter) when the

majority of sampling was conducted, differences between males and females were not

apparent and Hg concentrations were generally more stable at the index sites. However,

inter-annual variation may be high at certain sites, e.g. Corbett Brook where Hg

concentrations in late summer 2006 were ~0.45 ug g-1 and in late summer 2007 were

~0.15 ug g-1. These large, unexplained changes in Hg concentrations indicate the short-

term relevance of strider contamination levels relative to longer-lived species such as fish

and possibly Usnea spp.

Stable isotope ratios suggested no link between carbon sources and Hg concentrations in striders. The majority of strider populations exclusively use terrestrial energy (22 of 41 sites had 0% aquatic carbon; Chapter 3), and only in rare instances do striders derive the majority of their biomass from aquatic prey (7 of 41 sites had >50% aquatic carbon, Chapter 3). There was no relationship between % aquatic carbon in the strider diet and Hg concentrations (Fig 4.6a). The lack of a direct link between carbon source and Hg concentrations in striders contrasts to previous work in lakes, where animals connected to the pelagic zone have higher Hg concentrations than those that use littoral energy (Power et al. 2002, Kidd et al. 2003). In our study it was expected that animals foraging in streams on aquatic biofilm, which is a mixture of algae, fungi and bacteria, may be exposed to higher amounts of Hg due to methylation by sulfur-reducing

130

bacteria (Gilmour et al. 1992). However, this methylation of Hg requires anoxic

conditions (Gilmour et al. 1992) that were rarely encountered in the well-oxygenated streams of this study (minimum dissolved oxygen concentration for sites sampled in 2004

= 6.4 mg/L, TDJ, unpublished data). Striders that use terrestrial carbon can get that energy either from consumption of terrestrial insects or of aquatic insects that process terrestrial particulate organic matter. These latter two sources are likely low in Hg due to limited methylation in the terrestrial environment (Grigal 2002). Enhanced methylation

has been shown, however, to occur as a result of the flooding of terrestrial vegetation

(Rudd 1995), and the highest concentrations observed in this study were at a site (Clark

Brook) where strider Hg concentrations increased from 0.4 and 0.3 ug g-1 in 2006 to 2.1 and 1.8 ug g-1 in 2007 in females and males, respectively, possibly due to flooding of the

area upstream of the site by beavers in the latter year (TDJ, pers. obs.). At this site,

striders were feeding on aquatic prey (aquatic C = 79%), suggesting a link between

methyl Hg release from flooded vegetation and subsequent uptake by algae. Flooding has

been shown to cause a net increase in Hg production (Driscoll et al. 1998, Hall et al.

1998), and could therefore exert greater control over Hg concentrations in stream biota than all other factors examined here.

For those striders that were entirely connected to the terrestrial food source pathway (and thus not influenced by changing water chemistry across sites), trophic level explained significant variation in Hg concentrations across sites (Fig. 4.6b), likely reflecting biomagnification stemming from the consumption of larger insects that are positioned higher in the food chain (Cabana and Rasmussen 1994) or cannibalism. It could also be due to biogeochemical changes in N and Hg cycling due to flooding or

131

anoxia that could simultaneously influence δ15N and Hg. Water striders had a high

proportion of their total Hg as methyl Hg, not surprising given their status as obligate

predators and the enhanced biomagnification of methyl relative to inorganic mercury at

higher trophic levels (Mason et al. 1996). Earlier studies have shown that % methyl Hg

is related to position in the food chain for benthic invertebrates (e.g. 35-50% methyl Hg

in grazers-detritivores and 70-95% methyl Hg in predators, Tremblay et al. 1996).

Because methyl Hg is the more toxic form of Hg and subsequently of greater interest in

fish, wildlife, and human health studies (Boening 2000), a high proportion of total Hg as

methyl Hg in a sentinel species is considered a positive attribute.

The combination of environmental sentinels in this study was useful for

determining where follow-up work on Hg cycling may be warranted, as the two sentinels

provided different types of information on Hg availability to ecosystems. While Usnea

spp. provided a clear picture of Hg deposition near the power plant, water striders appeared more likely to reflect the complexities associated with Hg cycling within terrestrial and aquatic food webs. Spatial assessments of Hg contamination using water

striders as a sentinel will therefore require an appreciation of their ecological

characteristics (such as potential growth rates and trophic level) as well as variation in

their Hg concentrations over the course of the growing season. In terms of absolute

abundance, sampling striders in the fall provides the highest likelihood of capturing individuals in sufficient numbers to analyze for total Hg and methyl Hg, and to perform

other analyses including stable isotopes or other contaminants. Due to their smaller body

size, rapid growth rates, low Hg concentrations and limited availability, nymphs should

only be sampled in studies concerned with ontogenetic or seasonal changes in Hg

132

concentrations. Male striders present problems given their greater seasonal variability in

Hg concentrations, particularly in the spring. This leaves females as the best candidate for sampling given their more stable Hg concentrations over time and their larger body size.

Overall, striders appear to have limited utility as a sentinel for aquatic Hg contamination simply because the majority of their biomass is derived from the terrestrial environment and they have no secondary route of exposure to this contaminant (i.e. waterborne Hg, Hall et al. 1997). However, a strong connection to the terrestrial environment may lend them to a role in linking measured Hg concentrations with predicted atmospheric Hg deposition, allowing a better understanding of spatial and temporal trends in Hg contamination, as well as serving as indicators of Hg concentrations in other organisms that consume terrestrial insects such as brook trout

(Mookerji et al. 2004, Jardine et al. 2005). Also, their apparent ability to undergo rapid change in Hg concentrations (based on seasonal data) may make them useful as short- term indicators of Hg availability, although this would have to be tested through controlled experimentation. Understanding the poor correlation between Hg in striders and Hg in Usnea spp. will also require further examination to better model the relationship between Hg emissions, deposition, and resultant concentrations in aquatic biota.

4.5 Acknowledgements

The authors thank T. Arciszewski, P. Emerson, A. Fraser, S. Fraser, L. Giardi, K. Lippert,

A. Townsend, J. Penny, T. McMullen, C. Blanar, C. Poole, R. Keeler, P. Brett, T. Barrett,

133

R. Engelbertink, M. Sullivan, C. Ritchie, S. McWilliam, N. Swain, M. Sabean, M. Nasr,

D. Perkman, L. Sweeney, D. Banh, J. O’Keefe, C. Paton, M. Savoie, A. McGeachy, O.

Nwobu, E. Yumvihoze, L. Baker, B. Wyn, E. Campbell, and E. Belyea for field and lab assistance. Funding was provided by the O’Brien Humanitarian Trust, the NB Wildlife

Trust Fund, the NB Environmental Trust Fund, the Grand Lake Meadows Fund, and the

Natural Sciences and Engineering Research Council Discovery, Canada Graduate

Scholarship and Canada Research Chair programs. Comments from H. Swanson

improved the quality of this manuscript.

4.6 References

Al-Reasi, H.A., Ababneh, F.A., and Lean, D.R. 2007. Evaluating mercury

biomagnifications in fish from a tropical marine environment using stable isotopes

(δ13C and δ15N). Environmental Toxicology and Chemistry 26: 1572-1581.

Anderson, W.L., and Smith, K.E. 1977. Dynamics of mercury at coal-fired power plant

and adjacent cooling lake. Environmental Science and Technology 11: 75-80.

Andersen, N.M. 1990. Phylogeny and taxonomy of water striders, genus Aquarius

Schellenberg (Insecta, Hemiptera, Gerridae), with a new species from Australia.

Steenstrupia 16: 37-81.

Beeby, A. 2001. What do sentinels stand for? Environmental Pollution 112, 285-298.

Belzile, N., Chen, Y.-W., Gunn, J.M., Tong, J., Alarie, Y., Delonchamp, T., and Lang,

C.-Y. 2005. The effect of selenium on mercury assimilation by freshwater

organisms. Canadian Journal of Fisheries and Aquatic Sciences 63: 1-10.

134

Blackenhorn, W.U., and Fairbairn, D.J., 1995. Life history adaptation along a latitudinal

cline in the water strider Aquarius remigis (Heteroptera: Gerridae). Journal of

Evolutionary Biology 8, 21-41.

Boening, D.W. 2000. Ecological effects, transport, and fate of mercury: a general review.

Chemosphere 40, 1335-1351.

Bourque, C. P.-A. 1992. Sulfur dioxide absorption by spruce forests downwind from a

coal-fired generator. PhD Thesis, University of New Brunswick.

Bourque, C. P.-A., and Arp, P.A. 1994. Sulfur dioxide plume dispersion and subsequent

absorption by coniferous forests: Field measurements and model calculations.

Boundary-Layer Meteorology 71: 151-168

Braune, B.M., and Malone, B.J. 2006. Organochlorines and trace elements in upland

game birds harvested in Canada. Science of the Total Environment 363: 60-69.

Burgess, N.M., Evers, D.C., and Kaplan, J.D. 2005. Mercury and other contaminants in

common loons breeding in Atlantic Canada. Ecotoxicology 14: 241-252.

Cabana, G., and Rasmussen, J.B. 1994. Modelling food-chain structure and contaminant

bioaccumulation using stable nitrogen isotopes. Nature 372: 255-257.

Cai, Y., Tang, G., Jaffe, R., and Jones, R. 1997. Evaluation of some isolation methods for

organomercury determination in soil and fish samples by capillary gas

chromatography-atomic fluorescence spectrometry. International Journal of

Environmental and Analytical Chemistry 68: 331-345.

Cunningham, P.A., Smith, S.L., Tippett, J.P., and Greene, A. 1994. A national fish

consumption advisory database – a step towards consistency. Fisheries 19: 14-23.

135

Cuvin-Aralar, M.L.A., and Furness, R.W. 1991. Mercury and selenium interaction: a

review. Ecotoxicology and Environmental Safety 21: 348-364. de Freitas Rebelo, M., Reboucas do Amaral, M.C., and Pfeiffer, W.C., 2003. High Zn and

Cd accumulation in the oyster Crassostrea rhizophorae, and its relevance as a

sentinel species. Marine Pollution Bulletin 46: 1341-1358.

Driscoll, C.T., Holsapple, J., Schofield, C.L., and Munson, R. 1998. The chemistry and

transport of mercury in a small wetland in the Adirondack region of New York,

USA. Biogeochemistry 40: 137-146.

Edgerton, E.S., Hartsell, B.E., and Jansen, J.J. 2006. Mercury speciation in coal-fired

power plant plumes observed at three surface sites in the southeastern U.S.

Environmental Science and Technology 40: 4563-4570.

Environment Canada. 2005. National Pollutant Release Inventory.

www.ec.gc.ca/pdb/npri

Evers, D.C., Han, Y.-J., Driscoll, C.T., Kamman, N.C., Goodale, M.W., Fallon Lambert,

K., Holsen, T.M., Chen, C.Y., Clair, T.A., and Butler, T. 2007. Biological mercury

hotspots in the Northeastern United States and Southeastern Canada. BioScience

57: 1-15.

Fairbairn, D.J. 1985. Comparative ecology of Gerris remigis (Hemiptera: Gerridae) in

two habitats: a paradox of habitat choice. Canadian Journal of Zoology 63: 2594-

2603.

Fairbairn, D.J. 1986. Does alary polymorphism imply dispersal polymorphism in the

waterstrider, Gerris remigis? Ecological Entomology 11: 355-368.

136

Fairbairn, D.J. 1997. Allometry for sexual size dimorphism: Pattern and process in the

coevolution of body size in males and females. Annual Review of Ecology and

Systematics 28: 659-687.

Fairbairn, D.J., Vermette, R., Kapoor, N.N., and Zahiri, N. 2003. Functional morphology

of sexually selected genitalia in the water strider Aquarius remigis. Canadian

Journal of Zoology 81: 400-413.

Fox, G.A., Grasman, K.A., Hobson, K.A., Williams, K., Jeffrey, D., and Hanbridge, B.

2002. Contaminant residues in tissues of adult and prefledged herring gulls from the

Great Lakes in relation to diet in the early 1990s. Journal of Great Lakes Research

28: 643-663.

Galbraith, D.F., and Fernando, C.H. 1977. The life history of Gerris remigis

(Heteroptera: Gerridae) in a small stream in southern Ontario. Canadian

Entomologist 109: 221-228.

Garty, J. 2001. Biomonitoring atmospheric heavy metals with lichens: theory and

application. Critical Reviews in Plant Sciences 20, 309-371.

Gilmour, C.C., Henry, E.A., and Mitchell, R. 1992. Sulfate stimulation of mercury

methylation in fresh-water sediments. Environmental Science and Technology 26:

2281-2287.

Green, R.H. 2005. Marine coastal monitoring: Designing an effective offshore oil and gas

environmental effects monitoring program. In: Armsworthy, S.L., Cranford, P.J.,

Lee, K. (Eds). Offshore oil and gas environmental effects monitoring – Approaches

and Technologies. Batelle Press, pp. 373-398.

137

Grigal, D.F. 2002. Inputs and outputs of mercury from terrestrial watersheds: a review.

Environmental Reviews 10: 1-39.

Hall, B.D., Bodaly, R.A., Fudge, R.J.P., Rudd, J.W.M., and Rosenberg, D.M. 1997. Food

as the dominant pathway of methylmercury uptake by fish. Water, Air, and Soil

Pollution 100: 13-24.

Hall, B.D., Rosenberg, D.M., and Wiens, A.P. 1998. Methyl mercury in aquatic insects

from an experimental reservoir. Canadian Journal of Fisheries and Aquatic Sciences

55: 2036-2047.

Hammar, J., Larsson, P., and Klavins, M. 1993. Accumulation of persistent pollutants in

normal and dwarfed Arctic char (Salvelinus alpinus sp. complex). Canadian Journal

of Fisheries and Aquatic Sciences 50: 2574-2580.

Hammerschmidt, C.R., and Fitzgerald, W.F. 2005. Methylmercury in mosquitoes related

to atmospheric mercury deposition and contamination. Environmental Science and

Technology 39: 3034-3039.

Hammerschmidt, C.R., and Fitzgerald, W.F. 2006. Methylmercury in freshwater fish

linked to atmospheric mercury deposition. Environmental Science and Technology

40: 7764-7770.

Harding, K.M., Gowland, J.A., and Dillon, P.J. 2006. Mercury concentration in black

flies Simulium spp. (Diptera, Simuliidae) from soft-water streams in Ontario,

Canada. Environmental Pollution 143: 529-535.

Hintelmann, H., Harris, R., Heyes, A., Hurley, J.P., Kelly, C.A., Krabbenhoft, D.P.,

Lindberg, S., Rudd, J.W.M., Scott, K.J., and St.Louis, V.L. 2002. Reactivity and

mobility of new and old mercury deposition in a boreal forest ecosystem during the

138

first year of the METAALICUS study. Environmental Science and Technology 36:

5034-5040.

Hrabik, T.R., and Watras, C.J. 2002. Recent declines in mercury concentration in a

freshwater fishery: isolating the effects of de-acidification and decreased

atmospheric mercury deposition in Little Rock Lake. Science of the Total

Environment 297: 229-237.

Hylander, L.D., and Goodsite, M.E. 2006. Environmental costs of mercury pollution.

Science of the Total Environment 368: 352-370.

Interagency Monitoring of Protected Visual Environments. 1994. IMPROVE

monitoring update volume 3, number 2. Air Resource Specialists, Inc., Fort Collins,

CO.

Iwata, H., Masukawa, T., Kito, H., and Hayashi, M. 1982. Degradation of methylmercury

by selenium. Life Sciences 31: 859-866.