BEFORE THE ILLINOIS POLLUTION CONTROL BQ

,S IFFED

S OFFICE

IN THE MATTER OF

:

)

AUG

0 $ 20066

STATE OF

ILLINOIS

PROPOSED NEW 35 ILL. ADM. CODE 225

)

R06ftllution Control Board

CONTROL OF EMISSIONS FROM

)

(Rulemaking - Air)

LARGE COMBUSTION SOURCES (MERCURY)

)

NOTICE OF FILING

PLEASE TAKE NOTICE that the Environmental Law and Policy Center has filed the

attached MICHAEL MURRAY REFERENCES IN SUPPORT OF TESTIMONY

.

DATED: August 8, 2006

Environmental Law and Policy Center

35 E. Wacker Drive, Suite 1300

Chicago, Illinois 60601

312-673-6500

Faith Bagel (Reg. No. 62

685)

Counsel for Environmental Law and Policy Center

---

BEFORE THE ILLINOIS POLLUTION CONTROL B

C E 1 V E D

RK'S OFFICE

IN THE MATTER OF

:

)

AUG 0 5 2005

PROPOSED NEW 35 ILL. ADM. CODE 225

)

R06-2

STATE OF ILLINOIS

CONTROL OF EMISSIONS FROM

)

(Rulema?

ngIPAWtr

ol Board

LARGE COMBUSTION SOURCES (MERCURY)

)

MICHAEL MURRAY REFERENCES IN SUPPORT OF TESTIMONY

The following documents are references in support of the testimony of Michael Murray

which was filed in PCB R06-25 on July 24, 2006

.

DATED: August 8, 2006

Environmental Law and Policy Center

35 E. Wacker Drive, Suite 1300

Chicago, Illinois 60601

312-673-6500

Faith Bugel (Reg. No. 62556 5)

Counsel `or Environmental Law and Policy Center

---

Ecotoxicology, 14, 193 221, 2005

V 2005 Springer Science + Business Media, Inc. Manufactured in The Netherlands

.

Patterns and Interpretation of Mercury Exposure in Freshwater Avian

Communities in Northeastern North America

DAVID C. EVERS," NEIL M. BURGESS, LOUISE CHAMPOUX, 3 BART HOSKINS,4 ANDREW MAJOR,'

WING M. GOODALE,' ROBERT J. TAYLOR ,6 ROBERT POPPENGA ° AND THERESA DAIGLE'

'BioDiversity Research Institute, 19 Flaggy Meadow Rd., Gorham, ME, 04038, USA

2Canadian Wildlife Service, Environment Canada, 6 Bruce St ., Mt. Pearl, NL, Canada AIN 4T3

-'Canadian Wildlife Service, Environment Canada, Ste-Foy, Quebec, Canada GI V 4H5

4United States Environmental Protection Agency, II Technology Dr ., N. Chelmsford 01863, MA, USA

5U.S. Fish and Wildlife Service, Concord, NH, 03301, USA

6Texas A &M University, Trace Element Research Lab, College Station, TX, 77843, USA

7University of Pennsylvania, School of Veterinary Medicine, Kennett Square, PA, 19348, USA

Accepted 4 December 2004

Abstract. A large data set of over 4,700 records of avian mercury (Hg) levels in northeastern North

America was compiled and evaluated . As Hg emissions remain poorly regulated in the United States and

Canada, atmospheric deposition patterns and associated ecological responses continue to elicit interest by

landscape managers, conservation biologists, policy makers, and the general public . How avian Hg

exposure is interpreted greatly influences decision-making practices . The geographic extent and size of this

data set is valuable in understanding the factors that affect the exposure of Hg to birds . Featured are

differences found among tissues, major aquatic habitats and geographic areas, between age class and

gender, and among species. While Hg concentrations in egg and blood reflect short-term Hg exposure, Hg

concentrations in liver and feather provide insight into long-term Hg exposure . Blood is a particularly

important matrix for relating site-specific exposure to methylmercury (MeHg) . The level of McHg is

generally 5-10x greater in adults compared to nestlings . Age also influences MeHg bioaccumulation,

particularly for individuals where MeHg intake exceeds elimination . Gender is of interpretive concern when

evaluating Hg exposure for species exhibiting sexual dimorphism and niche partitioning . Based on two

indicator species, the belted kingfisher (Ceryle alcyon) and bald eagle (Haliaeetus leucocephalus), we found

McHg availability increased from marine, to estuarine and riverine systems, and was greatest in lake

habitats. A large sample of > 1,800 blood and egg Hg levels from the common loon

(Gavia immer)

facilitated a suitable comparison of geographic differences . Although some clusters of highly elevated Hg

exposure (i .e ., blood levels > 3 .0 gg/g, ww and egg levels > 1 .3 µg/g, ww) were associated with hydrological

and biogeochemical factors known to increase MeHg production and availability, others were not. Geo-

graphic areas without a relationship between Hg exposure and biogeochemical processes were associated

with emission or waterborne point sources. Differences in Hg exposure among species are primarily cor-

related with trophic position and availability of McHg . Although piscivorous species were repeatedly

'To

whom correspondence should be addressed

:

Tel.: 207-839-7600;

Fax :

207-839-7655 ;

E-mail: david.evers@briloon.org

---

194

Evers et al .

shown to have some of the highest MeHg levels of the 38 species analyzed, insectivorous birds in both

aquatic and terrestrial habitats (such as montane areas) were also found with elevated McHg levels . A

better understanding of the factors confounding interpretation of Hg exposure provides an effective basis

for choice of indicator species and tissues according to 12 selected scenarios. This and the national need for

spatiotemporal monitoring of McHg availability require careful consideration

of indicator species choice

.

Only then will local, regional, continental, and even global monitoring efforts

be effective .

Keywords : bird; loon; methylmercury; monitoring; indicator species

Introduction

The ecological impact from atmospheric deposition

of mercury (Hg) has emerged as a major global

environmental issue. Global concerns stem from the

broad geographic extent of contamination,

the

increasing global signal of Hg deposition, and, until

recently, a general lack of regulations to control

many uses and the disposal of Hg (United Nations

Environment Programme, 2003) . In North Amer-

ica. decades of increasing Hg deposition appear to

have reversed in some areas (Engstrom and Swain,

1997; Schuster et al ., 2002 ; Fevold et al ., 2003),

including the Northeast (Kamman and Engstrom,

2002), but the need to identify and monitor eco-

logical changes remains a high priority (Mason et

al ., 2005) . Federal, state and/or provincial regula-

tion of atmospheric mercury emissions in the Uni-

ted States and Canada is in place for some industrial

sectors (i.e ., municipal and medical waste incinera-

tion), but is currently lacking for others

(i.e ., coal-

fired electrical generators and mining)

.

Not all

environmental Hg is related to atmospheric depo-

sition. Many past and even current inputs of

waterborne Hg sources occur throughout North

America and the Northeast . These are related to

past improper waste disposal of Hg at weapons

facilities (Halbrook et al ., 1999), chlor-alkali plants

(Fimreite,

1974 ; Gardner et aL, 1978 ; Barr, 1986;

Adair et al ., 2003), mercury, gold, and silver mines

(Elbert and Anderson,

1998 ; Henny et al ., 2002 ;

Seiler et al ., 2004 ; Weech et al ., 2004) and govern-

mental storage facilities (Moore et al ., 1999) as well

as current inputs from wastewater treatment plants

(Glass et al ., 1990)

.

The U.S. Environmental

Protection Agency

(USEPA) investigated the ecological impacts of Hg

based on key wildlife species as a basis for potential

regulatory actions (USEPA 1997) . An outgrowth of

this effort was the development of a generic wildlife

criterion value for bird and mammal species (Nic-

hols and Bradbury, 1999) .Since the USEPA Report

to Congress (U SEPA 1997), scientific investigations

on the biogeochemical process of methylmercury

production and availability have dramatically im-

proved our basic knowledge (Morel et al

.,

1998 ;

Lucotte et al

., 1999 ; Wiener et al ., 2003)

. A better

understanding of the mechanisms of Hg transfer

and fate has improved the ability to predict meth-

ylmercury (MeHg) production and

availability

(USEPA 2002), particularly in freshwater habitats

of northeastern North America (Evers and Clair,

2005) . This has resulted in a greater insight into now

identifying specific geographic areas and biota at

greatest risk to Hg exposure and effects.

Birds are at particularly high risk to Hg toxicity

because many species are at high trophic levels

(e.g .,

susceptible to biomagnification), are long-

lived (e.g., susceptible to bioaccumulation), are

vulnerable to neurological and reproductive im-

pacts from elevated Hg levels, and are frequently

subjected to multiple anthropogenic stressors .

Using birds as bioindicators

of

MeHg availability

The use of piscivorous birds as bioindicators of

Melig availability and risk in freshwater systems is

common (e.g ., Fimreite, 1974 ; Barr, 1986 ; Scheu-

hammer, 1987 ; Wolfe et al ., 1998 ; Rumbold et al .,

2001 ; Henny et al ., 2002 ;

Evers et al .,

2003),

al-

though insectivorous birds are increasingly being

used as well (Wolfe and Norman, 1998 ;Gerrard and

St. Louis, 2001

; Adair et al ., 2003) . Historically, Hg

exposure was primarily determined by killing birds

and was therefore based

on organs analysis

(Thompson,

1996) . Although collection of viable

eggs continues to be a relevant lethal method widely

used (Braune et al ., 2001), non-lethal sampling ef-

forts based on blood (Bowerman et al

.,

2002 ;

---

Evers et al ., 1998; Fevold et al., 2003), feathers

(Burger, 1993), and abandoned eggs (Scheuhammer

et al., 2001 ; Evers et al., 2003) are increasingly a

more frequently used approach . Since Hg concen-

trations in different avian tissues reflect different

temporal scales of past Hg exposure, care must be

taken in considering Hg pharmacokinetics when

selecting the best avian tissue to match specific bi-

omonitoring objectives .

This paper represents a three-year effort through

the U.S. Department of Agriculture's

North-

eastern States Research Cooperative (NSRC) to

comprehensively compile and synthesize bird Hg

data across northeastern North America. The pa-

per's purpose is to describe this large data set and

use the information to identify . and assess

the

importance of factors that affect exposure

and

bioaccumulation of Hg.

Figure 1

.

Distribution of Hg sampling effort for all bird species, 1969-2003

.

Mercury exposure in Northeast North America

195

Methods

Source data sets

We targeted the collection of Hg data in birds

from aquatic freshwater systems in New England,

New York, and eastern Canada (eastern Ontario

to the Canadian Maritimes) (Fig . 1). The Great

Lakes and Lake Champlain were not included

within our data set. Only blood Hg data for belted

kingfishers and bald eagles were gathered

from

saltwater systems; these data were used to dem-

onstrate differences among major aquatic habitats

(Fig. 2). The majority of data (> 90%) were pro-

vided by BioDiversity Research Institute, Cana-

dian

Wildlife

Service,

U .S .

Environmental

Protection Agency, and the U .S. Fish and Wildlife

Service .

I

lb

omete+s

---

196

Evers et al .

All tissue data represent analysis of total Hg on

a wet weight (ww), in the case

of

feathers, fresh

weight

(fw), basis in µg/g (or ppm) . Estimated

values or ranges

of

the proportion

of

MeHg in a

particular tissue are cited

for each within the

Discussion

section. The term juvenile

means

young-of-year birds and adults signify individuals

at least one year

of

age. Latin names for those

species within our Hg dala sets arc provided in

Appendix I

.

Common loon blood and egg Hg sampling

locations were converted into an ESRI ArcView

point shapefile

(i.e .,

formatting georeferenced

parameters in a way that can be used by spatial

software). Egg Hg values were converted to adult

female blood equivalency with y = 1 .5544x +

0.2238

&tS

•4

lyllnvr GtidFtl4

O

Abet

C. Wmnn

o

LiriAn

o

emery

Figure 2.

Distribution of sampling effort by habitat type for the belted kingfisher and bald eagle in Maine .

(Evers et al

.

2003)

.

A six latitudinal minute by six

longitudinal minute polygon grid created in Coor-

dinate Grid Maker

2.29

was layered on the loon

data. The 6-min interval was chosen as the best

resolution to balance local and regional trends . The

loon Hg shapefile was spatially joined to the grid

polygon where the arithmetic mean of all the points

falling within a grid cell was calculated . These global

means were then displayed in 1 .0 pg/g (ww) inter-

vals .

Lahoratorv methods

The data utilized in this compilation were gener-

ated at a number

of

laboratories over a period

of

several years

.

Although there

were some

---

differences

in sample preparation and analytical

methods, all analyses included

quality control

(QC) samples to allow evaluation of accuracy and

precision, and all laboratories

utilized atomic

absorption spectroscopy to measure Hg concen-

trations .

Sample types collected and submitted to the

laboratories for analysis primarily included avian

blood, feathers, and eggs. Blood samples were

either in sealed capillary tubes or in glass or plastic

vacutainer-type collection

tubes. Samples that

were severely clotted were not analyzed unless the

entire sample could be removed from the collec-

tion tube .

Feather samples were either analyzed whole or

as subsamples following homogenization . Aliquots

of feathers were obtained by reducing individual

feathers to small pieces with either stainless steel

scissors or a Spex 6800 cryomill. '

Egg samples generally required homogeniza-

tion; a task that was sometimes complicated by the

egg samples that were fully formed. Egg samples

that were largely soft tissue were homogenized by

either a Tissuemiser or a small food processor/

blender prior to subsampling. Eggs containing

hard parts and feathers were homogenized with a

blender or with a Spex 6800 cryomill. Only loon

eggs were corrected for moisture loss

Most blood, feather, and egg samples required

digestion prior to analysis. This was accomplished

by following a procedure similar to EPA 245 .6, in

which nitric and sulfuric acids were used in con-

junction

with

potassium

permanganate

and

potassium persulfate to solubilize the tissue and

convert any bound Hg to the free Hg2+

ion

(Lobring and Potter, 1991). Prior to analysis, ex-

cess KMnO4

was reduced with hydroxylamine

hydrochloride and the samples were made to vol-

ume with deionized water

.

Analysis of digest solutions was based on the

"cold vapor" atomic absorption

spectroscopy

method

first introduced

by Hatch and

Ott

(1968). Using either a manual

or automated

approach, Hg` in solution was reduced to Hg °

with SnCl2, the Hg° was transferred to the gas

phase, and the Hg° -containing gas was swept

into an atomic absorption cell

.

Mercury levels

were determined by comparing sample

absor-

bance peak heights with those

of calibration

standards .

Mercury exposure in Northeast North America

197

A subset of samples was analyzed by a direct

determination method that did not require sample

digestion (EPA

7473)

(U .S .

EPA,

1998). A

homogenized, dry sample was placed in a tared

nickel boat, weighed, and then placed into a tube

furnace. A stream of 02 assisted in sample com-

bustion and carried free or organic-bound

Hg

species through a heated catalyst and onto a gold

trap where the free Hg° was collected . When the

sample had been combusted for a sufficient length

of time, the gold trap was heated and the released

Hg°

was carried

through a pair of atomic

absorption cells where it was measured .

This

method required samples that were particularly

well-homogenized because only a small sample

mass could be accommodated in the nickel boats

.

Each batch of samples processed and analyzed

was accompanied by a number of QC samples,

including a method blank, spiked blank, certified

reference material, duplicate sample, and spiked

sample. Typical detection limits for data presented

here were 0.0025 µg/g(ww) . Precision as measured

as relative percent difference of duplicate pairs was

approximately 85% and accuracy as measured by

recovery of certified reference materials and spiked

samples was 80% .

Statistical analysis

Mercury concentrations are expressed as arithme-

tic means with standard deviations (SD) in the

tables

and

geometric

means

with

variation

expressed as standard

error (SE) in figures

.

Arithmetic means and SD are provided for com-

parative

purposes

with

published

literature

.

Because sample sizes were regularly small and were

therefore not normally distributed,

statistical

analysis was conducted on the exponentiated value

of the mean of the log-transformed values. Log-

transformed data were normally distributed based

on normal probability plot residuals . Homosce-

dasticity was checked with Bartlett's test, which is

sensitive to the normality assumption . IMP soft-

ware (SAS Institute Inc., 2001) was used to per-

form statistical analysis. Hypotheses were tested

using one-way analysis of variance (ANOVA) .

Testing was followed by post-hoc tests using Tu-

key-Kramar honestly significant different (Tukey's

test) if the ANOVA demonstrated significant dif-

ferences (Zar, 1999). JMP's Tukey's test output

---

198

Evers et al .

did not include actual probability values and in-

stead indicated significance when numbers were

positive. Therefore, only probability values "less

than" and "greater than" 0 .05 are shown in the

Results section . Student's t-tests were used when

comparing paired data sets. A non-parametric test,

the Kruskal-Wallis One-Way ANOVA, was used

in some cases to compare multiple independent

groups. JMP software corrected for inequity of

unbalanced data sets . We used an alpha of 0 .05 for

our level of significance

.

Results

A total of 4,769 Hg concentrations representing 38

species and six tissue types are recorded within the

NSRC avian database (Appendix

1-3; Fig. I) .

Samples were collected between

1969 and 2003

with the majority (> 84%) from 1995 to 2003. Six

factors were identified as having significant influ-

ence on the interpretation of avian Hg levels . The

NSRC data set was used to demonstrate how these

factors influence Hg exposure

.

Influences of tissue type

Mercury data collections totaled 2,158 blood, 943

egg, 281 muscle, 1,100 feather, 239 liver, and 48

kidney samples

(Appendix

I) .

Approximate

respective inter-tissue comparative ratios based on

o Qaeeec

a New E, grane

•

Canadian Maallmea

678

Mu

e(5

9,

5

blood for common loons breeding in northeastern

North America were 0 .4:1 :2:6:15 (egg:blood :mu-

scle:feather:liver). For a site-specific subset of Hg

exposure data (south-central Quebec, New England

and Canadian Maritimes), there were no significant

geographical

differences

among

tissue

ratios

(p>0.05; with the exception of blood) (Fig . 3).

Muscle Hg levels among eight waterfowl species

were categorized by four major foraging guilds

during the breeding season and indicated significant

differences between piscivores versus each of the

other three foraging guilds and insectivores versus

herbivores (Fig. 4)

.

Intra-and inter-tissue relationships were stron-

gest in the following three pairings : (I) adult and

juvenile blood, (2) adult female blood and egg, and

(3) juvenile feather and blood. Data analyzed were

based on sampling efforts that represented pairings

from the same breeding territory (i.e ., each pair of

adult and juvenile blood Hg levels in tree swallows

was from the same nesting box). Paired adult-

juvenile blood Hg levels in common loons had a

significant relationship (r2=0.63, p < 0.01) as they

did for tree swallows

(r 2=0.74, p<0.01). Paired

adult female blood and egg Hg levels were sig-

nificantly related

in loons (r2=0.79, p<0.01)

(based on Evers et al., 2003) and tree swallows

(r2 =0.49, p<0.01). Paired eaglet feather and

blood

Hg

levels

were

significantly

related

(r2=0.67, p<0.01) (based on Welch, 1994) .

Figure 3 . Comparison of geometric mean

3

SE of Hg levels in four tissue types for adult common loons breeding in south-central

Quebce, New England, and the Canadian Maritimes (n = number of respective samples by region). The arithmetic mean t SD

of Hg in liver is used for comparative purposes with the literature . Liver Hg values in New England are from Pokras et al . (1992) .

---

Age as a factor

Five species of birds were used to demonstrate

relationships in blood Hg levels between juveniles

(< 2 months of age) and adults (> 1 year of age)

(Fig. 5). Two species were insectivores (tree swal-

low and song sparrow) and three species were pi-

scivores (common loon, common merganser, and

1 .21

d

0 .80

Z

060

0.21

0.00

Mercury exposure in Northeast North America

199

belted kingfisher) . While adult tree swallows had

significantly higher blood Hg levels than nestlings

(p<0.05), adult song sparrows did not have sig-

nificantly different blood Hg levels than their

fledged young (p > 0.05). Significantly higher blood

Hg levels in adults versus juveniles were found in

all three species of piscivorous birds (p < 0 .05) .

Ratios of adult-juvenile Hg levels were : song

Hertivwe(n=10)

Ornr9vure(n=e5)

Insectivore (n=39)

Pisdvore(n=51)

Figure 4

.

Comparison of geometric mean t SE of Hg levels in muscle among

foul foraging guilds of waterfowl (n

= number of

samples). Means not sharing a common letter are significantly different (p<0 .05). Waterfowl species represented by foraging guild

arc: herbivores - Canada goose; omnivores - mallard, American black duck, green-winged teal, and ring-necked duck ; insectivore

- common goldeneye; and piscivore - hooded merganser and common merganser

.

SOSP

(0-17,16)

TRES

BENZ

COME

(n..73.53)

(n-187.117)

(1 .69,11)

COi-O

to .452, 766)

Figure 5 .

Comparison of geometric mean + SE of blood Hg levels of five species for two age classes (n

= number of juvenile

samples, number of adult samples) . Means not sharing a common letter within a given species are significantly

different

(p<0.05).

Species codes arc : SOSP, song sparrow; TRES, tree swallow; BEKI, belted kingfisher; COME, common merganser and; COLO,

common loon

.

---

200

Evers ei a!.

sparrow

1 .7:1, common

merganser

3.6:1,

tree

swallow 5 .9:1, belted kingfisher 5 .6 :1, and common

loon

10.6 :1

.

Gender as a factor

Three

species

were

used

to

demonstrate

a

relationship in blood Hg levels between male and

female adults (Fig. 6) . Blood Hg levels in male

common loons were significantly higher than

females (p<0.01). In belted kingfishers and tree

swallows,

there was

no significant difference

between male and female blood Hg levels (p > 0.05)

.

Aquatic habitat comparisons

Data for the belted kingfisher and bald eagle

showed relationships in MeHg availability for four

major aquatic habitat categories: marine, estua-

rine, riverinc, and lake. Only individuals sampled

in Maine were used for an analysis of inter-habitat

differences (Fig

. 2) .

Both species exhibited an

increasing trend in which mean blood Hg levels in

marine <

estuarine

< riverine < lake (king-

fishers, Fig

. 7 and eagles, Fig

. 8) . Adult blood Hg

levels were significantly higher in kingfishers for-

aging on lakes and rivers versus marine habitats

(p < 0.05), but not different in rivers versus estu-

aries (p>0.05)

.

Eaglet blood Hg levels were

oFemale

® Male

TRES

0- 36, 12)

BEKI

pe 51 60)

significantly higher from nests along lakes versus

those adjacent to marine (p<0.05) and estuarine

(p < 0 .05) habitats. Blood Hg levels in eaglets from

nests along rivers were significantly higher than

those along marine habitats (p<0 .05) .

Geographic differences

Based on a standard species (common loon) and

tissue types (blood and egg) and a large sample size

(n = 1,882), spatial heterogeneity in McHg avail-

ability was demonstrated across

northeastern

North America (Fig. 9). The proportion of mean

blood and converted egg Hg concentrations (see

"Source data sets" in Methods section) within a

six latitudinal minute by six longitudinal minute

cell grid (n = 300 cell grids) was 19.7% for 0-I µg/

g, 45.6% for

1-2 fig/g, 20.7% for 2-3 pg/g,

8.7%

for 3-4 pg/g, and 5.3% for > 4 pg/g .

Variation in species

Three sites provided an opportunity to compare

Hg levels in multiple species within the same area

(which avoids confounding factors related to

geographical differences): two examples were on

lakes and one was on a river . On Aziscohos Lake,

eggs of five species (three piscivores, one insecti-

vore, and one herbivore) were compared (Fig . 10) .

COLO

tr 364.406)

Figure 6 .

Comparison

of

geometric mean

i

SE of

blood Hg levels

of

three species by sex

(n =

number

of

female samples, num-

ber

of

male samples)_ Means not sharing a common letter within a given species are significantly different (p

< 0.05).

Species codes

are :

TRES,

tree swallow ;

BEKI,

belted kingfisher and

;

COLO,

common

loon

.

---

a2

Marine

in =6.41

Estuary

in-31 .12)

Mercury exposure in Northeast North America

201

Rive,

In-26,16)

lake

(n-67,46)

Figure 7.

Comparison

of

geometric mean

± SE of

blood Hg levels for juvenile and adult belted kingfishers in Maine among four

major aquatic habitat types (n = number

of

juvenile samples, number

of

adult samples). Means not sharing a common letter arc

significantly different (p

< 0.05) ;

comparisons are case sensitive.

0.6

0.s

0.4

a

d

0

0.7

P

m

02

0 .1

0

Piscivore egg Hg levels were significantly higher

(p<0.05) than insectivores (tree swallows) and

herbivores (wood duck). Larger piscivores tended

to have higher egg Hg levels than smaller pisci-

votes

(i .e .,common loon > common merganser >

hooded merganser). Similar patterns were docu-

mented on Flagstaff Lake in egg Hg levels (com-

mon

loon

>

common

merganser

>

belted

kingfisher) (Fig. II).

Blood Hg levels were compared for I I insec-

tivorous birds from the Sudbury and Charles

be

Maine

in =30)

Estuary

in =7)

River

in

=6)

Lake

(n=34)

Figure 8 .

Comparison

of

geometric mean ±

SE of

blood Hg levels for juvenile bald eagles in Maine among four major aquatic

habitat types (n = number

of

samples). Means not sharing a common letter are significantly different (p

< 0.05) .

Rivers (Fig. 12). There were no significant differ-

ences in blood Hg levels between the rivers for four

of the most common passerines sampled at each

site (Kruskal-Wallis one-way analysis ofVariance ;

p = 0.65). Age classes were combined because

song sparrow blood Hg levels were not signifi-

cantly different between breeding adults and nes-

tlings

(Fig. 5) .

Generally,

blood

Hg

levels

increased with body weight . For seven selected

songbirds, there was a significant correlation be-

tween mean weight and blood Hg levels (r 2=0.72,

---

202

Evers et at.

Mean

Hg Level (ugi v.w)

-o-'o

0-20

,` 2

=30-40

M~40

2W

I

Kibme en

Figure 9. Geographic distribution of mercury levels in adult common loon blood and blood equivalents based on eggs, 1993-2003

.

p = 0.02, y =

0.0162x + 0.0054) . Granivorous

songbirds

such

as

the American

goldfinch

had

blood Hg levels significantly lower than insectiv-

orous songbirds

(p<0.05)

.

Discussion

Selecting the correct tissue

The pharmacokinetics of McHg and total Hg

(how both forms are distributed throughout the

body) are fairly well known in birds; their

understanding provide insights toward selecting

the correct tissue for meeting specific research or

monitoring

objectives. Ingestion

of

dietary

MeHg appears to be readily absorbed into the

blood (83% in common loons, Fournier et al .,

2002)

and is thereafter distributed to various

body tissues, namely the liver, kidney, brain,

spleen, and muscle. Some tissues, such as the

liver, are terminal endpoints where McHg is

largely unavailable for remobilization, whereas

MeHg deposited in muscle tissue is available and

remobilizes during feather molt. Blood and

feather Hg samples are generally taken nonle-

thally. Although blood can be taken immediately

after death through cardiac puncture (Henny

et al .,

2002), it cannot be taken through veni-

puncture after death for reliable comparison with

living bird blood Hg levels (because of rapid and

nonlinear moisture loss) .

Mercury concentrations in avian tissues can

indicate different (I) modes of toxic action,

(2)

MeHg and total Hg composition, (3) exposure

timeframes . and

(4) elimination abilities . Six tis-

sues routinely used for determining Hg exposure in

birds are described in our data set

.

---

Tissue interpretation

Blood is the best tissue for evaluating short-term

dietary uptake. Mercury in blood is primarily

MeHg (>95%) in both piscivores (Fournier et al .,

2002) and insectivores (Rimmer et al., 2005). The

half-life of MeHg in the blood of chicks under-

going feather molt was three days in common

B

.4

WOW

TRES

HEGU

(r . 1)

In-e)

(n=4)

Mercury exposure in Northeast North America 203

HOME

CO

COLO

(n-s)

(n

in

=43)

Figure 10 .

Comparison of geometric mean ± SE of egg Hg levels among species on Aziscohos Lake, Maine (n = number of

samples). Means not sharing a common letter are significantly different (p<005) . Species codes are: WODU, wood duck; TRES,

tree swallow; HEGU, herring gull ; HOME, hooded merganser ; COME, common merganser and; COLO, common loon .

Figure

11 .

Comparison of geometric mean + SE of egg Hg levels among species on Flagstaff Lake, Maine (n = number of sam-

ples). Means with different letters are significantly different (p<0 .05). Species codes are: WODU, wood duck; BEKI, belted king-

fisher; COGO, common goldeneye; COME, common merganser and; COLO, common loon .

loons (Fournier et al., 2002) and 5-6 days in

Cory's shearwaters (Monteiro and Furness, 2001) .

In non-molting adults, the half-life of MeHg in the

blood

was

greater :

for

Cory's

shearwater

(Calonectris diomedea) the half-life was 40-60 days

(Monteiro and Furness, 2001) and for the mallard

was 74 days (Heinz and Hoffman, 2004) . In our

study, adult loon blood Hg levels, which were

---

204

Evers et at.

wommoreminilli

0

P

IS)

U OF W E

-sl

)

I, 91 In

P

S SP

In

8

FL

SR

A

G

Figure 12 .

Comparison of geometric mean t SE for blood Hg levels (adult and juvenile combined) among species on or near the

Sudbury and Charles Rivers, Massachusetts (n = number of samples) . Means not sharing a common letter are significantly dif-

ferent (p<0-05). Species codes are: AMGO, American goldfinch; YWAR, yellow warbler; WODU, wood duck; GRCA, gray cat-

bird ; COPE, common ycllowthroat ; SOSP, song sparrow; SWSP, swamp sparrow; WIFL, willow flycatcher; TRES, tree swallow ;

RWBL, red-winged blackbird and ; NOWA, northern waterthrush,

collected 60-120 days post-arrival to the breeding

lake, were strongly correlated with prey fish Hg

levels and therefore primarily reflected uptake of

dietary Hg from the breeding lake (Fevold et al .,

2003; Evers et al., 2004, Burgess and Hobson, in

press; Champoux et al., in press)

.

As in blood, egg Hg levels are primarily in the

MeHg form: in loons it was >95% (Scheuhammer

et al., 2001 ; Evers et al ., 2004), in seabirds it is

>90% (Fimreite et al .,

1974). Because female

blood Hg levels are highly correlated with egg Hg

levels for the common loon (Evers et al., 2003),

eggs and their outer membranes are also pertinent

tissues for predicting Hg exposure within a bird's

breeding territory (Heinz and Hoffman,

2003) .

How predictive egg Fig levels are for the breeding

territory depends on amount of time the female

spent within the territory prior to egg-laying and

Hg body burden levels accumulated during the

winter and migration. Generally, piscivores arriv-

ing from marine overwintering areas have been

exposed to lower MeHg availability than levels

found on freshwater aquatic systems (Fig . 7, 8)

.

Rapid equilibrium of dietary MeHg uptake and

blood McHg levels thereafter plays an important

role in forming a strong McHg relationship be-

tween eggs and breeding-season blood levels . The

strength of this relationship is also impacted by

intraclutch differences in Hg levels. In common

loons, intraclutch variation between two eggs was

25% (Evers et al., 2004), in lands it averaged 39%

between the first and second eggs (Becker et al .,

1994), and in common merganser clutches of > 10

eggs we sometimes found variations of one order

of magnitude (although nest parasitism in cavity-

nesting ducks is common and it unknown if mer-

ganser clutches represent more than one female) . A

standardized comparison of egg Hg levels among

locations therefore requires knowledge

of egg

laying order

.

Virtually all Hg in a feather is MeHg (Thomp-

son and Furness, 1989) and is sequestered for long

time

periods

allowing

retrospective

analysis

(Frederick et al., 2004). Feather Hg reflects blood

Hg levels at the time of molt (Bearhop et al ., 2000),

however, if McHg is depurated in the muscle tissue

(as is the case for individual birds with a high

dietary McFlg uptake), it is available for remobi-

lization. Therefore, feather Hg levels reflect both

site-specific dietary uptake of MeHg and body

burden. Feather Hg generally reflects 70-93% of

the muscle MeHg burden (Burger, 1993); there-

fore, there can also be chronic bioaccumulation of

McFlg, particularly for highly exposed individuals

(i .e., where MeHg ingestion exceeds elimination)

.

This attribute makes feather Hg levels a relevant

tissue for evaluating chronic body burdens, par-

ticularly when considering the stability of MeHg in

the feathers (Appelquist et al ., 1984). However,

individual variation in physiological response to

---

Hg (Bearhop et al., 2000), as well as the broad

differences in inter-species pharmacokinetics,

re-

quires careful evaluation of risk

.

Four internal tissues are commonly used for Hg

exposure investigations: brain, liver, kidney and

muscle. Although MeHg crosses the blood-brain

barrier and can have significant impacts on brain

functions, brain tissue is best harvested from rel-

atively fresh carcasses and therefore is a more

difficult tissue to use in field studies . It is more

commonly used for analysis in experimental stud-

ies (Heinz, 1975; Finley et al., 1979; Scheuhammer,

1988) .

Liver is one of the more commonly analyzed

internal tissues for Hg in birds (Sundlof et al ., 1994 ;

Augspurger et al., 1998; Pokras et al., 1998; Cohen

et al., 2000). Liver and kidney filter toxins such as

MeHg and effectively demethylate MeHg using

selenium (Se) bonds that form a nontoxic Hg-Se

protein complex (Stoewsand et al ., 1974). Scheu-

hammer et al., (1998a) found a nearly 1 :1 molar

ratio of Hg:Se in the liver and kidney of common

loons and common mergansers . He also demon-

strated that the proportion of McHg in the liver

and kidney declined as total Hg concentrations

increased (i.e ., liver and kidney total Hg levels were

independent of MeHg concentrations). Therefore,

determining levels of currently toxic Hg in the liver

and kidney requires analysis of McHg concentra-

tions, while concentrations of Se-bound inorganic

Hg provide an indication of past MeHg exposure .

The "7:3:1 rule" is an often-used conversion

factor for liver (ww), feather (fw), and muscle (ww)

tissue Hg concentrations (Appelquist et al., 1985) .

Although Thompson et al., (1990) demonstrated

the weakness of this conversion approach, these

tissues in common loons of three different regions

in northeastern North America follow the "7:3 :1

rule" (Fig. 3) .

Unlike the liver and kidney, muscle Hg levels

generally have proportionally

higher levels of

MeHg (80-100%) in the common loon and com-

mon merganser (Scheuhammer et

al., 1998a) .

Muscle tissue Hg levels are generally examined in

waterfowl (Pearce et al ., 1976; Braune et al., 1999 ;

Cohen et al., 2000), in part, to determine potential

human health risks. Our data set supports other

findings that muscle Hg levels are generally less

than liver and kidney (Gardiner, 1972 ; Gochfeld,

1980) and that piscivorous waterfowl have muscle

Mercury exposure in Northeast North America

205

Hg levels significantly greater than other foraging

guilds of waterfowl species (Fig . 4) .

Age affects Hg exposure

A potential limitation of using birds as an indica-

tor of Hg is the inability to identify individual age

once breeding begins. The knowledge of a bird's

age is critical for evaluating Hg bioaccumulation .

Marking techniques, such as uniquely numbered

or colored leg bands, provide a reliable method for

tracking individual age and removing this limita-

tion. Such techniques have been used successfully

to evaluate time relationships with Fig. In cases

where

MeHg

ingestion

exceeds

elimination,

feather Hg levels increase with age (Evers et al .,

1998; Rimmer et al., 2005). Scenarios where indi-

viduals can depurate and demethylate ingested

McHg at a similar annual rate of ingestion lack a

positive correlation

with increasing age and

feather Hg .levels (Furness et al ., 1990; Thompson

et al., 1991; Gochfeld et al., 1996; Donaldson et

al., 1997; Fevold et al., 2003)

.

While bioaccumulation of McHg can be a crit-

ical interpretive factor related to age in situations

with high Hg exposure, there is also a common

pattern for adult blood Hg levels to exceed those

of unfledged juveniles in areas with even low Hg

exposure (Fig. 5). Burger and Gochfeld

(1997)

documented adult Franklin's gull (Larus pipixcan)

and herring gull blood Hg levels to be significantly

greater than juveniles from the same colony of the

same year. Differences in Hg levels between age

classes (juvenile vs. adult) is dictated by (1) stage of

juvenile feather molt and (2) partitioning of forage

base by size of potential prey .

A major deputation route for MeHg is via the

feather (Crewther et al .,

1965), and therefore

feathers are a useful indicator for monitoring Hg

body burdens (Burger,

1993) .

However, the

interpretation of feather Hg levels requires

an

understanding of feather-molt chronology (Fur-

ness et al., 1986). Blood and feather McHg levels

are highly correlated when

blood is sampled

during feather molt. Birds have the

ability to

rapidly transfer dietary uptake of MeHg

from

red blood cells to growing feathers

(Fournier

et al ., 2002; Kenow et al., 2003). The physio-

logical capacity of birds to process MeHg this

way appears to be great. Fournier et al., (2002)

---

206

Evers et al .

found

loon

chicks, experimentally

dosed with

MeHg concentrations substantially greater than

those found in nature, were still able to effec-

tively depurate much of the McHg into emerging

feathers . When juvenile feather molt ends, blood

MeHg levels thereafter increase (Spalding et al .,

2000a; Fournier et al ., 2002). This ability to

rapidly

transfer

blood

MeHg

into

growing

feathers partly accounts for the significant dif-

ference in blood Hg levels between adults and

juveniles prior to fledging. Our data indicate that

this is relevant for piscivorous and insectivorous

birds. For the song sparrow, blood Hg levels of

recently fledged young when compared with lo-

cally breeding adults, demonstrated no signifi-

cant difference between the age classes. However,

difference of age ratios among species (Fig . 5)

indicates Hg differences between age classes are

also dictated by other factors .

Generally, prey choice differs between adults

foraging for themselves versus for their young . For

example, the size of fish prey selected by juvenile

common loons increases as they grow larger (Barr,

1996) . We found adult blood Hg levels of piscivore

species were 3 .6 to 10 .6 times higher than those in

unfledged young (Fig. 5) and that variation sug-

gests that adults are foraging on prey that are larger

and have higher levels of Hg. In agreement, Burgess

and Hobson (in press) found that adult loons fed at

a higher trophic level (as indicated by stable-nitro-

gen isotope ratios) and had higher blood Hg levels

than did juvenile loons. These differences in blood

Hg and trophic levels were related to differences in

body weight in the loons. Blood sampling for belted

kingfishers and tree swallows typically occurs when

the adults are 6 times the weight of the juveniles,

while in common loons, adults are sampled when

they are usually 10 times the weight of juveniles

.

Conversely, common mergansers are usually sam-

pled when the adult females are relatively similar in

size with the young. Such rough correlative rela-

tionships across age classes between body weight

and prey-size selection are likely contributing to age

class differences in Hg exposure

.

Gender affects Hg exposure

Although most studies indicate differences in Hg

levels between male and female birds are not sig-

nificant (Burger, 1993; Burger, 1995; Gochfeld

et al., 1996), there are exceptions (Hoffman and

Curnow, 1979; Braune and Gaskin, 1987; Don-

aldson and Braune, 1999). When there are differ-

ences in Hg levels between sexes they can be

dictated by (I) deputation of Hg in eggs, (2) sexual

dimorphism, and (3) niche partitioning of the

forage base . Although deputation of McHg in eggs

is an important mechanism for elimination, deri-

vation of MeHg from serum proteins and a steady-

state equilibrium with dietary uptake of MeHg

likely compensates for the loss of MeHg from the

body burden within weeks (Furness and Green-

wood, 1993) or possibly days (Kambamandi-Di-

mou et al ., 1991). Differences in blood Hg levels

between sexes of adult loons sampled > 60 days

after eggs are laid regularly demonstrates other

factors are involved, Choice of prey items is likely

the primary factor dictating differences in Hg lev-

els between sexes. Levels of MeHg in prey items

vary according to species, trophic status, age, size,

and habitat associations (Wiener and Spry, 1996)

.

Based on comparisons of three species, the loon

exhibited significant differences of blood Hg

between sexes. Loon Hg levels were also greater in

liver for males vs . females (Pokras et al., 1998).

Common loons are sexually dimorphic. On aver-

age, males are 21% larger than females (Evers,

2005). The larger males apparently forage on larger

prey fish based on the correlative strengths between

favored prey fish Hg levels (yellow perch,

Perca

flavescens)

and their blood Hg levels (Evers et al .,

2004). Larger fish generally have higher Hg levels

than smaller fish of the same species from the same

location (Weiner and Spry, 1996; Drysdale et al ., in

press; Kamman et al., 2005). Therefore, blood Hg

levels are higher in male common loons because

they are foraging on larger fish .

In general, blood Hg differences between male

and female belted kingfishers are not significant,

however, within individual pairs Hg levels are

typically significantly different. Because males and

females are relatively similar in size (females tend

to be slightly heavier), prey size is likely not the

driving factor; rather males and females partition

foraging niches (Albano, 2000) .

Hg patterns in aquatic habitats

Aquatic systems are one of the more at-risk

ecosystems for McHg hioavailability because one

---

of the better-known methylating organisms, sul-

fate-reducing bacteria, inhabit this environment

(Gilmour et al., 1992). To adequately compare

MeHg bioavailability across four major aquatic

habitat types requires a standard

species, age

class and sampling tissue. Both the bald eagle

and belted kingfisher fit these criteria. A subset

of the blood Hg data from both

species in

Maine

indicates

interspecies

agreement

that

MeHg availability

increases

from marine

to

estuarine to riverine to lake ecosystems . Because

atmospheric deposition of Hg is relatively

uni-

form across the Maine study area (VanArsdale

et al., 2005), with some significant local excep-

tions, interpretation

of the hydrological

and

biogeochemical factors influencing Hg methyla-

tion and availability and their relationships with

bird blood Hg levels is presented

.

Although marine systems are well known for

their elevated biotic MeHg levels,

those levels

primarily represent long-lived species

with top

trophic status (e.g ., swordfish and shark species) .

A standard comparison between freshwater and

in-shore marine systems documents the latter has

significantly lower MeHg availability .

In-shore

marine systems appear to be more effective in

diluting

MeHg production

versus

freshwater

systems, although in-shore

habitats geochemi-

cally greatly vary in MeHg production (Ham-

merschmidt and Fitzgerald, 2004). Estuaries are

dynamic communities that are influenced by tidal

actions and varying volumes of fresh and salt

water. They are generally hydrologically hetero-

geneous landscapes that have less of an ability to

dilute

.Hg inputs than marine systems . Although

tidal exchanges do regularly provide an impor-

tant flushing mechanism

that lowers

McHg

availability (Lamborg et al

.,

In

Press),

tidal re-

sponses vary in magnitude daily,

weekly, and

monthly .

Methlymercury availability

therefore

varies tremendously within and between estuaries

(Shriver et al., 2002) .

Based on our analysis, the ability of inorganic

Hg to be converted to MeHg and become

available to biota is greater in freshwater versus

saltwater habitats. Gariboldi et al., (1998) also

documented prey items were higher in Hg from

freshwater versus saltwater

habitats based on

sampling efforts with the wood stork

(Mycteria

americana).

Riverine habitats tend to have higher

Mercury exposure in Northeast North America

207

levels of MeHg availability than estuaries,

but

tend to have lower levels when compared

to

lakes. Comparisons of MeHg

availability be-

tween adjoining riverine and lake habitats based

on crayfish, fish, and birds consistently

show

greater Hg exposure on lakes (Fimreite,

1974) .

Flushing abilities within riverine systems

are a

driving force for these differences

.

Therefore, in coastal regions, MeHg exposure

arising from atmospheric Hg deposition is gener-

ally greatest in piscivorous birds

foraging on

freshwater lakes. Lake hydrology

and biogeo-

chemistry largely determine the degree of aquatic

McHg exposure. Lakes with low pH (< 6.3, Meyer

et al ., 1995; Burgess and Hobson, in press), large

areas of scrub-shrub

and emergent wetlands

(Kramar et al., 2005), and large areas of exposed

shoreline substrate of organic or sandy soils that

are frequently inundated and dried through the

summer and fall

(i .e., reservoirs; Evers and Re-

aman, 1998) are predictive of elevated blood Hg

levels in the common loon . Although newly cre-

ated reservoirs are well known for their ability to

enhance

McHg

production

and

availability

through the decomposition of vegetation (Jackson,

1988; Lucotte et al., 1999; Gerrard and St. Louis,

2001), this phenomenon is generally viewed as

short-lived (i .e ., < 10 years in secondary consum-

ers) (Lucotte et al., 1999). Some reservoirs have

longer lasting abilities to enhance Hg methylation

and have the potential to be some of the highest

risk aquatic habitats (Evers et al., 2004) .

Geographic differences exist

There are continental patterns in the availability of

MeHg. Long-term sampling efforts across North

America indicate a significant west to east trend

exists with northeastern North America exhibiting

the highest levels (Evers et al ., 1998, 2003; Scheu-

hammer et al., 2001). Significant within-region

differences are primarily driven by hydrological

and biogeochemical

factors and point source

influences. The collection of > 1,800 blood and egg

Hg levels for the common loon across New Eng-

land, New York and eastern Canada enabled us to

effectively evaluate

areas of greatest concern

(Fig. 9). Clusters of elevated MeHg availability

were found in the western Adirondack Mountains

of New

York,

north-central

Massachusetts,

---

208

Evers et al.

southeastern New Hampshire, western mountains

of Maine, and a small area cast of Orrington,

Maine in the United States (Fig . 9) . In Canada,

areas with high MeHg exposure were in eastern

Ontario, south-central Quebec and southern Nova

Scotia. There did not appear to be smooth spatial

trends in loon Hg levels across northeastern North

America, as highly elevated Hg levels were

scattered among low Fig concentrations in almost

every region sampled

.

In the Adirondack Mountains and eastern

Canada, clusters of elevated MeHg availability

were likely related primarily to lake acidification

.

It is well established that lakes with low pH con-

tain fish with higher levels of Hg than same-size

and species of fish in lakes with more circumneu-

tral pH levels (Wiener et al ., 1990; Winfrey and

Rudd, 1990; Drysdale et al., in press). Oligotroh-

phic lakes in eastern Canada and parts of New

England and New York arc susceptible to in-

creased rates of anthropogenically derived sulphur

deposition

(i.e., acid rain) (Driscoll et al ., 2001) .

Although there is evidence of declining levels of

atmospheric input of sulfur dioxide, base cation

levels are lowered in many systems where re-

sponses in lake pH levels are lagging behind pre-

dictive models. Therefore, these acidified lakes

continue to be a cause of concern for their ability

to enhance McHg productivity. Associations be-

tween lake acidity, fish Hg levels and lower com-

mon loon productivity have been observed in the

U.S. and Canada (Meyer et al., 1998; Burgess et

al., 1998a). Scheuhammer and Blancher (1994)

predicted up to 30% of lakes in central Ontario

have the potential to adversely impact common

loon productivity

.

Other areas of concern are related to topog-

raphy and lake hydrology

(e.g ., western moun-

tains of Maine) and point sources (both airborne

and waterborne). Airborne sources in southern

New England appear to contribute to greater-

than-expected loon Hg levels in southeastern

New Hampshire (Evers, 2001) and Orrington,

Maine .

Waterborne

point

sources

are

well

known in eastern Massachusetts, such as on the

Sudbury River. There, investigations associated

with the Nyanza Superfund Site have docu-

mented associated Hg contamination >25 km

downstream from the point source (Wiener and

Shields, 2000) .

Species Hg exposure patterns

Differences in Hg levels among species are dictated

by trophic level and availability of MeHg

(i.e.,

aquatic vs. terrestrial and marine vs. freshwater

;

low exposure vs. high exposure). Trophic structure

is a primary driver of variability in MeHg bio-

magnification (Cabana et al., 1994). Recent evi-

dence indicates that the trophic status of an

aquatic vertebrate is based primarily on the com-

plexity, both longitudinal and vertical, of the

planktivorous community (Chen et al., 2005) .

Methylmercury can biomagnify several orders of

magnitude in aquatic ecosystems. For example, the

average bioconcentration factor for the common

loon in Maine lakes is 1 .37 x 106

(based on

unfiltered water for total Hg) (Evers et al ., 2004) .

The

degree

of

McHg

biomagnification

through aquatic-based food webs is the primary

reason for the multitude of Hg studies on obli-

gate piscivores. Particular emphasis has been

placed on larger species for which trophic status

is potentially greatest; such species include the

common loon (Meyer et al., 1995, 1998; Burgess

et al .,

1998a,

b ;

Evers et al .,

1998, 2003 ;

Scheuhammer et al., 1998b, 2001 ; Fevold et al .,

2003; Burgess and Hobson, in press; Champoux

et

al .,

in

press),

bald

eagle

(Grier,

1974 ;

Wiemeyer et al ., 1984; Frenzel and Anthony,

1989; Bowerman et al., 1994; Anthony et al .,

1999; Bowerman et al., 2002), osprey (Cahill et

al., 1998; DesGranges et al., 1998), wading birds

(Gariboldi et al., 1998; Bouton et al., 1999;

Spalding et al ., 2000b; Henny et al., 2002), and

seabirds (Braune, 1987; Burger and Gochfeld,

1995; Monteiro and Furness, 1995, 1997; Braune

et al ., 2001). Other foodweb pathways important

for MeHg transfer are generally of lesser concern

because trophic status of the endpoint species is

generally lower than piscivores . Benthic-based

MeHg transfer through bivalves has been investi-

gated using various diving species of waterfowl

(Ohlendorf et al., 1986; Henny et al., 1991 ; Braune

et al., 1999; Cohen et al., 2000; Wayland et al .,

2002), while such transfer through macroinverte-

brates (larval and adults) (Bishop et al ., 1995 ;

Wolfe and Norman, 1998; Reynolds et al ., 2001 ;

Gerrard and St. Louis 2002, Adair et al., 2003) and

vegetation has also been described (Fimreite, 1974 ;

Langis et al., 1999) .

---

Elevated MeHg bioavailability

in specific ter-

restrial

ecosystems

within

northeastern

North

America has recently been documented. Montane

environments without standing water appear to

have the ability to generate MeHg . Rimmer et al .,

(2005) documented Bicknell's Thrush blood Hg

concentrations for 21 mountain locations (arith-

metic mean of 0. 14 f 0.08 Vg/g, ww with a range of

< 0.01 to 0.70 µg/g, ww) at levels similar to those

found in many of the insectivorous songbirds sam-

pled along rivers in Massachusetts (Fig. 12) .

The comparison of multiple species within the

same area and habitats, while using appropriate

tissues and minimizing confounding factors (such as

age class and sex), is the optimal approach for

determining interspecies relationships of Hg expo-

sure. Based on such past studies (Dustman et al .,

1972; Fimreite, 1974; Langis et al ., 1999) and our

data sets, Hg exposure can be approximately pre-

dicted by foraging guilds. An all-purpose ranking

from low to high Hg exposure for birds is : terrestrial

herbivores, aquatic herbivores, terrestrial insecti-

vores, benthivore-bivalves,

benthivore-macroin-

vertebrates, small piscivores, and large piscivores .

Table 1 .

Summary of recommended avian bioindicators, age/sex class, and tissue type for 12 scenarios in freshwater, estuarine, and

terrestrial systems in northeastern North America

Mercury exposure in Northeast North America

209

Exposure of Hg in scavengers and omnivores is

broad and dependent on opportunistic food sources

(Fimreite, 1974). Our ranking assumes MeHg

availability is driven by atmospheric deposition and

is not universal in application, because some habi-

tats such as montane ones contain insectivorous

birds that have Hg exposure greater or equivalent to

piscivores .

Recommended bioindicators

We recommend species and tissue types that best

indicate 12 targeted scenarios (Table 1) based on

the analysis of our data set, the recommendations

made by the working group (USEPA Hg Mason

et al., 2005), species' ubiquitous within northeast-

ern North America,

and logistical feasibility.

Identified indicator species are not universal and

may be only relevant to the scenario posed. Many

of our chosen bioindicators are also useful for

determining McHg effects through such endpoints

as long-term reproductive success. For example,

bald eagle breeding populations

are used in

Michigan (Bowerman et al., 2002) and common

Juvenile = unfledged young which have yet to reach completion of feather molt and fledged young = young-of-the-year that have

completed feather molt

.

2 When using egg tissue from these species, only use composite values for entire clutch to avoid wide intra-clutch variation

.

Scenario

Species

Tissue type

Comparison of mailer aquatic habitat types

Belted kingfisher

Adult & fledged young

Blood & egg'

Bald eagle

Juvenile

Blood & feather

Lake > 25 ha

Common loon

Adult

Blood & egg

Common merganser

Adult female

Blood & egg'

Lake 125 ha

Common loon

Juvenile

Blood

Hooded merganser

Adult female

Blood

River

Common merganser

Adult female

Blood & egg 2

Belted kingfisher

Adult & juvenile

Blood& egg'

Tree swallow

Adult & juvenile

Blood & egg

Estuaries

Ammodramus

sparrow spp .

Adult & fledged young

Blood

Emergent wetlands

American bittern

Adult

Blood & egg

Virginia rail

Adult

Blood & egg

Song sparrow

Adult & fledged young

Blood

Red-winged blackbird

Adult & fledged young

Blood & egg

Shrub-scrub wetlands

Waterthrush spp .

Adult

Blood

Swamp and Song sparrows

Adult

Blood

Montane areas

Bicknell's thrush

Adult & fledged young

Blood

Deciduous forest

Wood thrush

Adult & fledged young

Blood

Coniferous forest

Catharus

thrush spp.

Adult and fledged young

Blood

Long-term risk in lakes

Common loon

Adult

Feather

Greatest risk in aquatic systems

Bald eagle

Adult

feather

---

210

Evers et a!.

loon

breeding

populations

with

color-marked

individuals

are

monitored

throughout

New

England (Evers et al., 2004) and Wisconsin (Meyer

et al., 1998; Fevold et al ., 2003) .

Our selections are species- and genera-specific

for illustrative purposes, but species with similar

foraging requirements, behavior, and natural

history patterns may be suitable surrogates

;

preferably, trophic status is similar. Evolving

techniques

in

stable

isotope

analysis

offer

numerous applications to matching trophic sta-

tus through analysis of tissues

(e.g .,

blood,

feather, egg, muscle, bone) . Such techniques

provide quantitative measures of trophic position

(Hobson, 1993; Bearhop et al ., 2000; Nisbet et

al., 2002; Dominguez et al ., 2003), dietary

emphasis

(i .e .,

freshwater vs. marine [Hobson,

1990; Mizutani et al., 1990; Bearhop et al .,

1999], marine versus terrestrial [Hobson, 1987 ;

Hobson and Scaly, 1991 ; Schmutz and Hobson,

1998]), contaminant bioaccumulation (Cabana

and Rasmussen, 1994; Kidd, 1998; Atwell et al .,

1998), and nutrient allocation to reproduction

(Hobson et al ., 1997; Hobson et al., 2000). Here

in, scenarios and associated avian selections re-

late primarily to freshwater breeding habitats

.

Selecting a standard species across multiple

aquatic habitats, particularly between freshwater

and saltwater ones, is difficult. The belted king-

fisher is a ubiquitous species that is an obligate

piscivore in all major aquatic habitats . As a

burrow-nesting species, repeated access to young

and adults is feasible for the, kingfisher. Sam-

pling efforts to determine site-specific exposure

should focus on blood. Intraclutch variability in

egg Hg levels appears to he high

(Lane et al .,

2004). Bald eagle pairs also commonly forage

within all major aquatic habitats. Adults are

difficult to capture, therefore, chicks are gener-

ally sampled to determine Hg exposure (Bower-

man et al., 2002). Blood and breast feathers are

the most common sampling tissues. Adult eagle

feathers from the nest site can be useful for

determining Hg exposure (Bowerntan et al .,

1994) and may reflect some of the highest Hg

levels within an aquatic ecosystem . However,

foraging habits of breeding pairs vary dramati-

cally within and between breeding seasons, hab-

itat type, and geographic area (Knight et al .,

1990; Kozie and Anderson, 1991 ; Anthony et al .,

1999). Although either fish or birds can comprise

the majority of prey remains at eagle nests,

Dominguez et al ., (2003) found that stable-

nitrogen isotope ratios showed little difference in

trophic status among nests in Newfoundland

.

The common loon is one of the better bioindi-

cators of lake-specific MeHg availability as it has a

top trophic position in the aquatic food web, is

long-lived, and in most cases remains within its

breeding territory for 4-6 months. Adult blood and

egg Hg levels reflect dietary Hg exposure of breed-

ing loons on lakes >25 ha . Territorial pairs occu-

pying lakes < 25 ha generally maintain and feed on

more than one lake

(i.e .,

multi-lake territories)

(Piper et al ., 1997). Because adult common loons

with multi-lake territories rarely bring food items

back to their natal lake to feed their young, blood

Hg levels of juvenile loons best represent McHg

availability on their natal lake. Common mergan-

sers are also obligate piscivores that can reflect lake-

specific MeHg availability. Sampling efforts for

adult females can be facilitated through the use of

artificial boxes. Similar to kingfishers, intraclutch

variation in egg Hg levels is high for mergansers and

dictates individual egg or composite analysis (ver-

sus selecting one egg). Other high-trophic level pi-

scivores that generally forage on lakes are not

optimal lake-specific indicator candidates because

(1) of their tendency to commonly use multiple

waterbodies within their breeding territory, (2) they

are difficult to efficiently capture and sample (e.g .,

great

blue heron), and (3) they regularly prey on

lower-trophic-level organisms such as benthic-

feeding fish (e.g ., osprey) and terrestrial birds and

mammals (e.g ., bald eagle). Instead, such species

best represent MeHg availability at a watershed

level . Double-crested cormorants

phalacrocorax-

aurltux

may be good indicators of multiple large

lakes and other aquatic systems .

Determining mercury exposure in riverine

habitats is most promising with the common

merganser, hooded merganser, belted kingfisher,

and

tree swallow. The belted kingfisher is

increasingly being used as an indicator for

assessing Hg in riverine systems (Baron et al .,

1997; Moore et al., 1999). Use of artificial

nesting boxes on riverine habitat and experi-

mental design interests for both piscivorous and

insectivorous birds can be achieved with the

hooded merganser and tree swallow .

---

Although

most

investigations

of

avian Hg

exposure have focused on waterbodies, wetlands

and

strictly

terrestrial

habitats

are

increasingly

being

included

during

risk

assessments .

In

emergent wetlands,

insectivores

best

reflect

MeHg

availability. Larger-bodied

insectivorous

birds

have greater

Hg

exposure

than

their

smaller counterparts (Fig .

12) . The Virginia

rail

(Rallus

limico/a)

is

a good

indicator

candidate

because

it

is more insectivorous than

the

sora

(Porzana carolina) and is more common and less

limited by marsh size than the American bittern

(Botaurus lentiginosus)

.

Clapper

rails

(Ra//us

longirostris) in San Francisco Bay had greater

body burdens of Hg than associated piscivorous

birds, such as terns

(U.S .

Fish and Wildlife

Service, 2003). The red-winged

blackbird had

some of the highest blood Hg levels of songbirds

within

a

Massachusetts

riverine

wetland

(Fig. 12). Although the red-winged blackbird and

sparrow species

(i.e ., song and swamp) are gra-

nivores most of the year, during the breeding

season they are obligate insectivores

.

Based on

limited Hg data, both the Louisiana

(Seiurus

motacilla)

and northern waterthrush

may be

insectivorous passerines at greatest risk in river-

me habitats; waterthrushes forage specifically on

aquatic organisms. Other relatively large-bodied,

insectivorous passerines associated with northern

aquatic systems, such as

the rusty blackbird

(Euphagus carolinus), may also be at risk; specific

Hg-sensitive habitats are acidic headwater areas

(Bank et al ., 2005) draining recently

logged

coniferous catchments (Porvari et al ., 2003) .

Methylmercury availability in terrestrial insec-

tivorous passerines is relatively unknown but a

recent compelling study by Rimmer et al . (2005)

indicates further investigations are needed. That

study documented Hg levels in Bicknell's thrush

and further comparison of the blood Hg level

ranges show an overlap with those of eaglets

;

thereby indicating equivalent trophic status of a

terrestrial-based insectivore with an aquatic-based

piscivore .

Recent investigations have demon-

strated that MeHg is present in foliage (approxi-

mately 1% of the total Hg content) (St

.

Louis

et al., 2001; Ericksen et al., 2003) . Miller et al .,

(2005) estimated MeHg availability to terrestrial

food webs using forest

foliage and modeled

deposition and concentrations of leaf,

litterfall,

Mercury exposure in Northeast North America

211

precipitation (wet and dry), and particulate Hg in

northeastern North America. Litterfall total Hg

concentrations from these models

were signifi-

cantly correlated with the blood Hg levels of Bic-

knell's thrush (Rimmer et al ., 2005). Conceivably

MeHg in litterfall and contributions of foliar total

Hg to saturated soils where potential methylation

environments exist are providing an important

basis for biomagnification of MeHg in inverte-

brates. Acidified environments further enhance

methylation (Furutani and Rudd, 1980 ; Xun et al .,

1987), and with the influence of heavy wet depo-

sition of acid ions

(i.e .,

acid rain), northeastern

North America's landscape

is generally more

acidic than pre-industrial times (Driscoll et al .,

2001). Soil acidification may impact bird popula-

tions in several ways (Graveland, 1998) including

the depletion of

soil

calcium levels. Breeding birds

have high demands of calcium for eggshell for-

mation and proper juvenile growth. The wide-

spread

depletion

of

environmental

calcium

availability in northeastern North America is now

linked to adverse effects on the distribution

of

wood thrush (Hylocichla mustelina) (Hames et al .,

2002). Unfortunately, the strong

link between

environmental acidification with MeHg produc-

tion and calcium depletion may be creating a sce-

nario where their synergy has the potential for

long-term, landscape-level impacts on insectivo-

rous passerine populations across much of north-

eastern North America .

Acknowledgements

We thank the many scientists representing agen-

cies, universities, industry, and organizations for

their data contributions . Several biologists at Bio-

Diversity Research Institute were instrumental in

collecting many of the tissue samples; they include,

Cory Counard, Chris DeSorbo, Joseph Kaplan,

Oksana Lane, James Paruk, Lucas Savoy, Kate

Taylor, Keren Tischler, and Dave Yates. Pierre-

Yves Daoust of the Atlantic Veterinary College

provided all muscle, kidney, and liver tissues ana-

lyzed for Hg by the Canadian Wildlife Service

.

Hydro-Quebec contributed otherwise unavailable

data. Gary Heinz and Stan Wiemeyer graciously

provided rapid reviews that improved the manu-

script. The U .S. Department

of Agriculture's

Northeastern

States

Research

Cooperative

---

Appendix 1

.

Basic geographic information, sampling time period, sample size by tissue and data source

Species

Latin Name

Project

Geographic

Area

Geographic Extent

Geographic

Extent

- Longitude

Sampling

Number of samples by tissue

Duration

-

Latitude

Min

Max

Min

Max

Begin End Blood

Egg Feather Liver Kidney Muscle

Common Loon

G,, e,

arm,

ME, MA,

42350277

51

.999444

-79

.447777

-60

.817

1986

2003 A

: 770

660

A

:

631

A

: 30 A

: 10

A

: 18

NH

. NY,

1: 452

J: 52

J

: 8

J

: 2

J

: 6

Black-crowned

Nycticorax

VT, NB

. NS

.

ON, PE, PQ

ME, PQ

43 487777

48

.460277

-74577777

-6853

1991

2003 J

: 17

J

: 6

Night-Heron

Great Blue Heron

ncctieorar

Ardeo herodias

ME

. MA, PQ

45

.030277

49

.460277

-74

.597222

-64

.722222

1991

2003 A

. 1

2

3.

. 8

1

: 17

Canada Goose

Brama bernicla

ME, NB, NL,

46

.133

57

.283

-66

.867

-54

.167

1989

1999

A

: 24

A

:7

A

: 15

Wood Duck

.4a sponsa

PE, PQ

ME, MA, NH,

42

.338333

49

.440555

-77

.480277

-66

.450

1990

2003 A

: 13

16

A

: 2

A

: 2

A

: 4

Mallard

Arias

NB, PQ

ME, NH,NY,

43

.693888

52

.917

-79

.411944

-63

.400

1990

2003 J

: 3

A

: 11

A

: 8

A

: 16'

platyrhynchos

NB, NL,

NS, PQ

NB, NL, NS,

43

.783

54

.167

-78

.830

-53

.250

1988

1993

A

: 39

A

:39

A

:62'

American Black Duck

Arias rubripes

PE, PQ

NB, NL,

45

.867

54167

-78

.838333

-53

.500

1990

1992 A

: I

A

: 8

A

: 8

A

: 24'

Green-winged Teal

Arias crecca

Ring-necked Duck

Arthya

collar,,

PE, PQ

NH, NB,

45

.500

51

.994722

-79

.280833

-55

.083

1990

1997

A

: 34

A

: 34

A

:39'

NL, PQ

1:I

P

.1

Common Goldeneye

Bueephala

ME

. NB, NL

.

43

.633

53167

-79

.083888

-56

.083

1989

2003 A

:6

22

A

:32

A

: 32

A

:39

Hooded Merganser

clangula

Lophodyte

NS, PQ

ME

. MA,

42

.26

53

.333

-79

.024166

-57

.833

1989

2003 A

: 13

45

A

: 11

A

. 17

A

:20

cucullatus

NH

. NB, NL,

1: 5

Common Merganser

Mergus

PQME, NH, NY

.

43,5825

54

.167

-79182222

-53

.500

1989

2003 A

: II

26

.A

: 24

A

: 31

merganser

VT, NB,

1: 69

1: 32

Osprey

pandion

NL, NS, PQ

ME, NH,PQ,

42

.968055

53

.333

-78

.752222

-60

.433

1989

2003 A

:2

23

A

: 10

A

: 12 A

: 4

A

: 6

haliaetus

NS, NB, NL, PE

1

: 58

J: 62

1

: 11

1

: 9

1

: 1

Bald Eagle

Ha/iaeetus

ME, NH, NB,

43

.836388

47

.000

-71

.611388

-59

.920

.

1969

2003 1

: 108

38

A

: 9

A

: 28 A

: 7

A

:8

leucocepha/us

NS, PEI

J: 2

J

: 2

J

: 2

American

Scolopax minor

ME, MA

44

.496666

44

.496666

-69

.237222

-69

.237222

2000

2002 J

:4

1

: 4

Woodcock

Herring Gull

Larus argentatus

ME

44

.76128

45

.05178

-71

.013383

-70

.8182

1990

2003 A

: 1

8

A

: I

1

: 15

1

: 1

---

*(Mallard, Am

. Black Duck, Am

. Green-winged Teal, and Ring-necked Duck muscle tissues represent composites

of adults and young-of-the-year)

.

Appendix

1. Continued

Species

Latin Name

Project

Geographic

Geographic Extent

Geographic

Extent

-

Longitude

Sampling

Duration

Number of samples by tissue

-

Latitude

Area

Min

Max

Min

Max

Begin End

Blood

Egg Feather Liver

Kidney Muscle

Belted Kingfisher

Ceryle

alcyon

ME, MA,

42

.172777

46

.024166

-73

.224166

-68

.27444

1997

2003 A

: 117

16

A

: 71

NH, VT

J: 183

J

: 13

Downy

Picoides

pubescens

MA

42

.347350

42

.347350

-71

.382380

-71

.382980

2003

2003 A

: I

Woodpecker

Willow Flycatcher

Empidonar

traillii

MA

42

.176240

42

.390130

71

.382380

-71

.319020

2003

2003 A

: I

1: 5

Great Crested

Myiorchus

crinitus

MA

42

.28558

42

.28558

-71

.44901

-71 44901

2003

2003 1

:4

Flycatcher

Eastern Kingbird

Tyrannus tyrannus

MA

42

.17386

42

.364408

-71

.31916

-71

.37542

2003

2003

26

Common Raven

Corvus corm

ME

44

.97933

44

.97933

-71

.0215

-71

.0215

2000

2000 J

: I

Tree Swallow

Tachycinem

bicolor

ME, MA

42

.17325

44

.94138

-71

.5831

-70

.9279

2000

2003 A

:53

55

J

: 73

Cliff Swallow

Petrochelidon

ME

44

.93642

45

.15257

-71

.0336

-70

.4466

1999

2001 A

: 19

pyrrhonom

U

: 6

Barn Swallow

Hirundo

rusrica

ME

44

.97047

44

.97047

-70

.7166

-70

.7166

1999

2001 A

: 3

Eastern Titmouse

Baeolophus bicolor

MA

42

.347350

42

.347350

-71

.382380

-71

.382380

2003

2003 J

: 1

Eastern Bluebird

Sialia

sialis

MA

42

.30914

42

.38378

-71

.49121

-71

.38748

2003

2003 J

:7

Bicknell's Thrush

Catharsis bicknelli

ME, VT

45

.183611

45

.183611

-70

.264722

-70

.264722

1999

2003 A

: 21

A

: 18

Gray Catbird

Dumetella

MA

42

.176240

42

.390130

-71

.382380

-71

.319020

2003

2003 A

: 2

carolinensis

1

: 4

Yellow Warbler

Dendroica petechia

MA

42

.176240

42

.390130

-71

.382380

-71

.319020

2003

2003 A

: 4

J

: 5

Northern

Seiurus

MA

42

.176240

42

.347350

-7t

.382380

-71

.319020

2003

2003 A

:2

Waterthrush

Common

noveboracensis

Geothlypis thrichas

MA

42

.176240

42

.390130

-71

.382380

-71

.319020

2003

2003 A

: 4

Yellowthroat

J

: 5

Song Sparrow

Melospiza Melodic

ME, MA

42

.176240

42

.390130

-71

.382380

-71

.319020

2001

2003 A

: 16

J

: 18

Swamp Sparrow

Melospiza

georgiana MA

42

.176240

42

.390130

-71

.382380

-71

.319020

2003

2003 A

: 5

J: 13

Agelaius

phoeniceus

MA

Red-winged

42

.175555

42351388

-71

.381111

-71

.321666

2003

2003 A

: 3

Blackbird

J: 3

Common Grackle

Quiscalus

quiscula

MA

42

.34631

42

.35972

-71

.37398

-71

.36956

2003

2003

6

Brown-headed

Molothrus

ater

MA

42

.176240

42

.176240

-71

.319020

-71

.319020

2003

2003 J

: I

Cowbird

American

C

.rd

.eli

trust

.

ME, MA

42

.176240

43

.731683

-71

.382380

-70

.566450

2000

2003 A

: 12

Goldfinch

---

Appendix 2

.

Tissue Hg levels (arithmetic mean ± SD and range) from nonlethal sampling efforts

Species

common loon

black-crowned night-heron

great blue heron

Canada goose

wood duck

mallard

American black duck

American green-winged teal

ring-necked duck

common goldeneye

common merganser

hooded merganser

osprey

bald eagle

American woodcock

herring gull

belted kingfisher

willow flycatcher

great crested flycatcher

eastern kingbird

tree swallow

cliff swallow

barn swallow

eastern bluebird

Bicknell's thrush

gray catbird

yellow warbler

northern waterthrush

common yellowthroat

song sparrow

swamp sparrow

red-winged blackbird

common grackle

American goldfinch

Blood (ww) Mean

-

SD (Range)

0

.01 -0

.01(<0

.01-0

.03)

Egg (ww) Mean ± SD (Range) Feather (fw) Mean ± SD (Range)

0

.78 a 0

.60 (0

.01

-

9

.00)

0.09 a 0

.04(0

.05-0

.12)

0

.12 a 0

.23 (0

.01 -0_94)

0

.33 a 0

.18 (0

.09

-

0

.72)

1.43

-

0

.86 (0

.28

-

3

.93)

0

.64

-

0

.44 (0

.15

-

1.90)

0

.19 s 0

.09(006-0

.38)

0

.45

-

0

.29 (0

.03

-

1.29)

0

.63 ± 0

.55 (0

.01

-

1

.63)

0

.56

± 0

.77 (0

.03

-

3

.03)

0

.12 t 0

.04 (0

.04 -

0

.21)

0

.19 a 0

.11 (0

.04

-

0

.64)

0

.04

-

0

.03 (0

.01

-

0

.07)

Adult12

.7

-

6

.6 (2

.2-63

.4)

5

.4

-

4

.8 (0

.3-25

.7)

0

.3 x 0

.1

(0

.3-0

.8)

1.6

-

0.3

(1,4-

1.8)

0

.9

a

0.5(0

.3-h8)

1.8

-

1.2

(0,6-

6

.6)

13

}

1 4

(0

.3-4

.5)

1.5

-

0.8 (0

.3

-

3.0)

2

.8 1 13(0

.8-70)

8

.0±4

.1(3

.3-17

.7)

10

.4 a 3

.8

(2

.7-

18

.0)

15

.6

-

13

.6 (0

.1 -38

.3)

14

.0

-

6

.8 (2

.8-24

.8)

I8

.1 a 15

.1 (4

.4

-

57

.0)

7

.2

-

7

.6 (0

.6

-

46

.1)

Juvenile

5

.2-2

.0(1

.3-6

.9)

8

.8 = 5

.4 (3

.3

-

31

.4)

8

.2-5

.6(0

.1-26

.5)

0

.2

-

0

.1 (0

.1

-

0

.3)

8.0-5

.4(3

.8-19

.8)

Adult

Juvenile

2.04

-

1.39

(0

.05-8

.63)

0

.27 ± 0

.34

(0

.01

-

358)

0

.28

-

0

.13

(0

.11

-

0

.52)

0

.49 ± 0

.56

(0

.03

-

1.76)

0

.05

±

0

.04

(0

.01 -0

.14)

0

.05 ± 0

.03

(0

.03

-

0

.08)

0

.21

-

0

.06

(0

.15-0

.31)

1.57

-

0

.59

(0

.74-2

.35)

0

.60 t 0

.47

(0

.03

-

2

.29)

0

.88

-

0

.55

(0

.07- 1

.91)

0

.68

- 0

.30

(0

.34- 1

.13)

1-42 ± 0

.18

(1

.29- 1

.54)

0

.31 1 0

.20

(0

.03

-

0

.81)

0.30

-

0

.27

(0

.01

-

1.20)

0.03 ± 0

.01

(0

.02- 0

.04)

0.46 ± 0

.13

(0

.28

-

0

.72)

0

.99 a 0

.82

(0

.07-4

.57)

0.17-0

.18(0

.01-1

.35)

0.43 a 0

.25 (0

.20-0

.80)

0.09 ± 0

.02 (0

.07-0

.11)

0

.41 = 0

.21

(0

.11

-

1

.00)

0.07 ± 0

.03 (0

.02-0,16)

0

.22 = 0

.10

(0

.08-0

.47)

0

.13 a 0

.03

(0

.11

-

0

.15)

0

.01

-

0

.01 (0

.01

-

0.02)

0

.29 a 0

.26

(0

.05 -0

.80)

0

.13 a 0

.07

(0

.08-0

.19)

0

.05

-

0

.02 (0

.03

-

0

.07)

0

.04 a 0

.03

(0

.01

-

0

.07)

0

.04

-

0

.03 (0

.01 -0

.08)

0

.92 -

0

.95

(015- 159)

0

.28

a

0

.13

(0

.15-0

.44)

0.10 x 0

.05 (0

.04-0

.17)

0

.35

-

0

.30

(0

.08

-

134)

0

.21

-

0

.14 (0

.01

-

0

.56)

0

.74 a 0

.47

(0

.22- 1

.45)

0

.30

a

0

.14 (0

.07

-

0

.48)

0

.67 a 0

.71

(0

.20- 1

.49)

0

.90 a 0

.38 (0

.46 -

1.13)

---

Appendix 3 . Tissue Hg levels (arithmetic mean 1 SD and range) from lethal sampling efforts

Species

common loon

- adult

common loon -juvenile

black-crowned night-heron

great blue heron

Canada goose

wood duck

mallard

American black duck

American green-winged teal

ring-necked duck

common goldencyc

common merganser

hooded merganser

osprey - adult

osprey -juvenile

bald eagle - adult

bald eagle - juvenile

herring gull

Liver (ww)

Mean ± SD (Range)

29.7 3 38.5

(1 .9- 154.0)

14.9 ± 25.6 (0 .3 - 92.6)

0.1 3 0.3 (0.1 - 0.2)

0.3 3 < 0 .1 (0.3-0.3)

0.5 3 0.2 (0 .1 - 0.8)

0.6 ± 0.4 (0.1 - 2.0)

0.8 ± 0.3 (0.2 -

1 .3)

0.6 ± 0.3 (0.2 -

1.5)

1 .5 i 1.8 (0 .1 - 8 .2)

4.7±3.0(11-12 .2)

10.6 t 11.2(0 .6-23 .0)

0.9 3 0.4 (0.2 -

1 .5)

2.2 3 2.2(0.6-11 .8)

1.2 3 0.9 (0.6 -

1 .9)

3.6 1 2 .5 (1.0-8 .2)

provided funding to compile and synthesize the

Hg databases .

References

Adair, B.M., Reynolds, K.D., McMurry, S.T. and Cobb, G .P .

(2003). Mercury occurrence in prothonotary warblers

(Protonolaria cinea) inhabiting a national priorities list site

and reference areas in southern Alabama

. Arch. Environ .

Conmm . Toxicol. 44, 265-71

.

Albano, D. (2000). A behavioral ecology of the Belted King-

fisher (Ceryle alcyon). Unpubl. PhD Thesis,

Univ. Mass .,

Amherst, Mass .

Anthony,

R .G .,

Miles, A.K.,

Estes, J.A. and Isaacs, F. B .

(1999). Productivity, diets, and environmental contaminants

in nesting bald eagles from the Aleutian Archipelago

.

Environ. Toxicol. Chem. 18, 2054-62

.

Appelquist, H., Asbirk, S. and Drahaek, I. (1984). Mercury

monitoring: Mercury stability in bird feathers . Mar. Pollut .

Bull. 15, 22-24 .

Appelquist, H., Draback, 1. and Asbirk, S . (1985). Variation in

mercury content of guillemot feathers over 150 years

. Mar.

Pollut Bull. 16, 244-48.

Atwell, L., Hobson K., A. and Welch, H . (1998). Biomagnifi-

cation and bioaccumulation of mercury in an arctic marine

food web: Insights from stable nitrogen isotope analysis

.

Can. J. Aqua± Sci. 55, 1114-21 .

Augspurger, T., Franson,

1.C., Converse, K.A ., Spitzer, P.R

.

and Miller, E.A . (1998). An epizootic of common loons in

coastal waters of North Carolina: Concentrations of

elemental contaminants and results of necropsies

. Environ

.

Toxic

.]. Chem. 17, 205-09 .

Mercury exposure in Northeast North America 215

15.7 t 14.7 (2.3-36 .7)

1.1 1 0.6 (0 .3

- 2.)

9.2 1

12.4 (0 .7 - 33.4)

0.6 ± 0.5 (0 .3 -

1 .0)

0.13 + 0.08 (0.02

- 0.26)

0.16 ± 0.12 (0.03 - 0.45)

0.21 ± 0.15 (0.02 - 0.59)

0.17 ± 0.11 (0.03-0.47)

0.33 + 0.20 (0.05-0.88)

1 .71 ± 1

.71 (0.08-6.77)

0 .96 ± 0.35 (0.04- 1.92)

1 .5 + 1.5 (0.6-4.3)

0 .1

0.430.4(0.2-1 .5)

0.3 3 0 .1

(0.2-0.3)

1 .59 + 1.32 (0.35 - 3.98)

Bank, M.S ., Loftin, C.S. and lung, RE. (2005). Mercury bio-

accumulation in two-lined salamanders from streams in

northeastern United States

. Ecotoxicalogy 14, 181-92 .

Baron, L.A ., Ashwood, T .I., Sample, B.E. and Welsh, C

.

(1997). Monitoring bioaccumulation of contaminants in the

belted kingfisher (Ceryle alcyon). Environ. Monitor. Assess .

47, 153-65 .

Barr, J .F. (1986). Population dynamics of the common loon

(Gavia immer) associated with mercury- contaminated wa-

ters in northwestern Ontario. Can . Wildl. Sem Occar. Pap .

56, Ottawa, Ontario, Canada .

Barr, J.F. (1996). Aspects of common loon

(Gavia immer)

feeding biology on its breeding ground

. Hydrohiologia 32,

119-44

.

Bearhop, S., Waldron, S., Thompson, D. and Furness, R

.

(2000). Bioamplification of mercury in great skua Catha-

racta skua chicks: The influence of trophic status as deter-

mined by stable isotope signatures of blood and feathers

.

Marine Pol/ut Bull. 40, 181-85.

Bearhop, S., Thompson, D.R ., Waldron, S., Russell,

I .C .,

Alexander, G. and Furness, R.W. (1999). Stable isotopes

indicate the extent of freshwater feeding by cormorants

Phalacrocarax carho shot at inland fisheries in England. J.

App!. Ecol. 36, 75-84.

Becker, P.H, Herring, D. and Furness, R.W. (1994). Differ-

ences in mercury contamination and elimination during

feather development in gull and tern broods . Arch. Environ .

Contain. Toxicol. 27, 162-67 .

Bishop, C.A ., Koster, M.D., Click, A.A ., Hassell, D .J.T. and

Jock, K. (1995). Chlorinated hydrocarbons and mercury in

sediments, red-winged blackbirds (Agelaius phoeniceu .s) and

tree swallows (Tachycineta bicolor) from wetlands in the

Great Lakes-St. Lawrence River basin

.

Environ. Toxic,,!.

Chem . 14, 491-501

.

Kidney (ww)

Mean ± SD (Range)

Muscle (ww)

Mean 1 SD (Range)

39.9 ± 25.1 (1.7-79.1)

4 .10 ± 3.40 (0.40 - 15.90)

35.9 t 24.9(11 .0-60 .7)

0 .90 3 0.81

(0.20 - 2.40)

1 .5 1 0.3

(1 .2-

1 .9)

1 .9 ± 0.9 (0 .4-3.3)

0.04 ± 0.03 (0.02-0.14)

---

216

Every et al .

Bouton, S.N ., Frederick, P. C., Spalding, M.G. and McGill, H .

(1999). Effects of chronic, low concentrations of dietary

methylmereury on the behavior of juvenile great egrets

.

Environ. Tax. Chem. 19, 1934-39

.

Bowerman,W.W., Evans, CD., Giesy, J . P. and Postupalsky, S

.

(1994). Using feathers to assess risk of mercury and sele-

nium to bald eagle reproduction in the Great Lakes Region

.

Arch. Environ. Cannot,. Taxirol

. 27. 294-98 .

Bowerman, W.W ., Roe, A . S., Gilhertson, M .1 ., Best, D.A.,

Sikarskie, J.G ., Mitchell, R.S. and Summer, C.L. (2002)

.

Using bald eagles to indicate the health of the Great Lakes'

environment . Lakes Revere. Res. Manage. 7, 183-87.

Braune, B. M. (1987). Comparison of total mercury levels in

relation to diet and moll for nine species of marine birds .

Arch. Environ . Cantam. To viral. 20, 217-24 .

Braune, B.M. and Gaskin, D.E. (1987). Mercury levels in

Bonaparte's gull (Larus philadelphia) during autumn molt in

the Quoddy region, New Brunswick, Canada

. Arch. Envi-

ron_ Con tam. Tocical . 16. 53'9-49 .

Braune, B .M

., Malone, R.J ., Burgess, N.M .,

Elliott,

1.E ., Gar-

rity, N., Hawkings, J., Hines, 1 .. Marshall, H., Marshall,

W.K ., Rodrigue, l., Wakeford, B., Wayland, M., Weseloh,

1) V. and Whitehead, P E . (1999). Chemical residues in

waterfowl and gamebirds harvested in Canada, 1987-1995

.

(',in . Wilsll Serv.Tech, Relic. Se, 326, Hull, Quebec, Canada .

Bounce, B.M

., Donaldson, G .M . and Hobson, K.A. (2001) .

Contaminant residues in seabird eggs from the Canadian

Arctic. Part I . Temporal trends 1975-1988 . Environ. Patent .

114, 39-54.

Burger, 1 (1993). Metals in avian feathers : Bioindicalors of

environmental pollution. Rev_ Environ . Taxirol. 5, 203-311 .

Burger, J . (1995). Heavy metal, selenium levels in feathers of

herring gulls (Larus argentatus) :

Differences due to year,

gender and age al Capture, Long Island

. Environ. Monitor .

Assess. 38, 37--50.

Burger, 1. and Gochfcld, M . (1995). Biomonitoring of heavy

metals in the Pacific Basin using avian feathers . Environ

.

Toxir,L Chem. 14, 1233-39

.

Burger, 1I and Goehfeld, M . (1997). Age differences in metals in

the blood of herring (Larus urgentatus) and Franklin's (Larus

ptplxmn) gulls . Arch . Environ. Cantam. Toxicol. 33, 436-40

.

Burgess, N. M. and Hobson, K.A. (2005). Bioaccumulation of

mercury in yellow perch and common loons in relation to

lake chemistry in Atlantic Canada

. Hydrobiologia . (in press) .

Burgess, N.M ., Evers, D.C ., Kaplan, 1.D ., Duggan, M. and

Kerekes, J.J (I 998a). Mercury and reproductive success of

common loons breeding in the Maritimes. In N.M. Burgess

led). Mercury in Atlantic Canada: A progress report

., pp

.

104-09. New Brunswick: Environment Canada, Sackville.

Burgess, N M., Evers, D . C. and Kaplan, 1.D. (1998b). Mercury

levels in the blood of common loons breeding in the Mari-

times and their prey. pp. 96-100. In N.M. Burgess (ed .) .

Mercury in Atlantic Canada: A progress report, pp. 96-100.

Environment Canada, Sackville, New Brunswick .

Cabana, GG

and Rasmussen 1, B. (1994). Modeling food chain

structure and contaminant hioaccumulation using stable

nitrogen isotopes

. Nature 372, 255-57.

Cabana, G., Tremblay, A., Kalff,

1I

and Rasmussen, J.R

.

(1994). Pelagic food chain structure in Ontario lakes : A

determinant of mercury levels in lake trout

(Salvelinus no-

maycush). Can. J. Fish. Aquae. Sci. 51, 381-89,

Cahill, T.M., Anderson, D.W ., Elbert, R.A ., Perley, B.P, and

Johnson, D R. (1998). Elemental profiles in feather samples

from a mercury-contaminated lake in central California

.

Arch_ Environ. Cantam. Taxirol. 35, 75-81

.

Champoux, L., Masse, D., Evers, D.C., Lane, 0., Planle, M .,

and Timmerman, S. Assessment of mercury exposure and

potential effects in common Loons in Quebec . Hydra-

hlologia (in press)

.

Chen, C.Y ., Stemberger, R.S ., Kamman, N.C., Mayes, B, and

Foil, C. (2005). Patterns of mercury biorecycrida Lion and

transfer in aquatic food webs across multi-lake studies in the

northeast U . S. Eroeoxiralogy 14, 135A8

.

Cohen, 1 B , Barclay, J .S ., Major. AR. and Fisher, 1 . P. (2000).

Wintering greater scaup as hiomonitors of metal contain,

..lion in federal wildlife refuges in the Long Island Region.

Arch. Environ . Centaur Tosirol . 38, 83-92.

Crewlher, W.G ., Fraser, RD., Leanox, F.G . and Lindley, H .

(1965). The chemistry of kcratins. In C . B. Anifmsen, M .L .

Anson, J.T. Edsoll and F.M. Richards (eds)

. Advances in

Protein Chemistry,

pp. 191-303. New York: Academic

Press

.

DesGranges, 1 . L., Rodrigue, 1., Tardif,

BB

and Laperle, M .

(1998). Mercury accumulation and hiomagnification in os-

prey (Pandion haliaems) in the James Bay and Hudson Bay

regions of Quebec_ Arch. Environ. Content. Taxienl.35 . 330

41

.

Dominguez, L., Monleveechi, W A., Burgess, N.M ., Brazil, 1L

and Hobson, K A . (2003). Reproductive success, environ-

mental contaminants, and trophic status of nesting bald

eagles in eastern Newfoundland, Canada_ J. Raptor Res

. 37,

209- I8

.

Donaldson, G.M . and Braunc, B M . (1999). Sex-related levels

of selenium, heavy metals, and organochlorine compounds

in American white pelicans (Pelecanas

erythrorhynco.s) .

Arch. Environ. Cantam. Taxirol. 37, 110-14 .

Donaldson, G.M

.,

Braune,

B.M

., Gaston, A.J. and Noble,

D.G. (1997). Organodtlorine and heavy metal residues in

breast muscle of known-age thick-killed murres (Uria tom-

via) from the Canadian Arctic. Arch. Environ_ Contain

.

To eicol. 33, 430-35

.

Disco[[, C.T.,

Lawrence, G.B .,

Bulger, A .1.,

Butler, T.J

.,

Cronan, C S., Eagar, C., Lambert, K.F.,

Likens, G.E .,

Stoddard, JL and Weathers, K.C. (2001). Acidic deposi-

tion in the northeastern United States: Sources and inputs,

ecosystem effects, and management strategies

. Biosrience

51, 180-98

.

Drysdale, C., Burgess, N M., Entermont, A ., Carter, 1. and

Brun, G. (2005). Mercury in brook trout, white perch and

yellow perch in Kejimkujik National Park and National

Historic Site. In A Renez and N. O'Driscoll (ads). Mercury

cycling in a wetland dominatedecosystem: a multidisciplinary,

study ., Pensacola Florida : SETAC Press (in press).

Dustman, E.H ., Uickel, L . F. and Elder, J.B. (1972). Mer .

cury in wild animals, Lake St. Clair, 1970. In R. Hartung

and RD Dinman (eds) . Environmental mercury contami-

nation_, pp. 46 52. Ann Arbor Michigan : Ann Arbor Sci

.

Publ .

---

Elbert, R.A. and Anderson, D.W. (1998). Mercury levels,

reproduction, and hematology in western grebes from

three California takes, USA

.

Environ.

Toxicol. Chem. 17,

210-13 .

Engstrom D, R. and Swain B, B. (1997). Recent declines in

atmospheric mercury deposition in the upper Midwest

.

Environ. Sc, Teahnol . 31,

960-67 .

Ericksen, 1., Gestic, M.S., Schorran, D., Johnson, D ., Lind-

berg, S. and Coleman, J . (2003). Accumulation of atmo-

spheric mercury in forest foliage .

Atm. Environ. 37,

1613-22 .

Evers, D.C. (2001). Assessing the potential impacts of methyl-

mercury on the Common Loon in southern New Hamp-

shire. Unpubl. report BRI-2001-2004 submitted to the New

Hampshire Dept

. Environ. Protection. BioDiversity Res .

Inst., Gorham, Maine

.

Evers, D.C. (2005)

.

Status assessment and conservation plan for

the common loon in North America

.

Final Draft, U .S. Fish

Wildl. Serv. Hadley, Massachusetts .

Evers, D.C. and Clair, T.A. (eds). (2005). Mercury in North-

castcrn North America: A synthesis of existing databases

.

Ecotoxicology

14, 7-13 .

Evers, D.C. and Reaman, P . (1998). A comparison of mercury

exposure and risk between artificial impoundments and

natural lakes measured in Common Loons and their prey,

1996-1997. Unpubl. report BRI 1998-2003 submitted to

Central Maine Power Co . BioDiversity Res . Inst., Gorham,

Maine

.

Evers, D.C., Lane, O.P., Savoy, L. and Goodale, W. (2004)

.

Assessing the impacts of methylmercury on piscivorous

wildlife using a wildlife criterion value based on the Com-

mon Loon, 1998-2003. Unpubl. report BRI 2004-2005

submitted to the Maine Department of Environmental

Protection .

BioDiversity

Research

Institute, Gorham,

Maine .

Evers, D.C., Taylor, K.M ., Major, A., Taylor, R .J ., Poppenga,

R.H. and Scheuhammer, A .M . (2003). Common Loon eggs

as indicators of methylmercury availability in North

America

.

Ecotoxicology 12,

69-81

.

Evers, D.C., Kaplan, J.D., Meyer, M .W., Reaman, P .S ., Bra-

selton, W.E ., Major, A., Burgess, N. and Scheuhammer,

A.M. (1998). A geographic trend in mercury measured in

common loon feather and blood

.

Environ . Toxicol. Chem .

17, 173-83

.

Fevold, B .M ., Meyer, M.W ., Rasmussen, P.W. and Temple,

S.A. (2003). Bioaccumulation patterns and temporal trends

of mercury exposure in Wisconsin common loons

.

Ecotox-

icology 12,

83-93 .

Fimreite, N. (1974). Mercury contamination of aquatic birds in

northwestern Ontario

.

J. Wild!

. Manage

.

38, 120-31

.

Fimreite, N., Brun, E., Froslie, A., Frederichsen, P. and Gun-

dersen, N. (1974). Mercury in eggs of Norwegian seabirds

.

Asiatic

1, 71-75

.

Finley, M.T ., Stickel, W.H. and Christensen, R.E. (1979). Mer-

cury residues in tissues of dead and surviving birds fed

methylmercury

.

Bull

. Environ. Contain. Toxicol. 21,

105-10 .

Frederick, P.C ., Hylton, B., Heath, J.A . and Spalding, M.G .

(2004). A historical record of mercury contamination in

southern Florida (USA) as inferred from avian feather tis-

sue

.

Environ

. Toxicol .

Chem. 23, 1474-78 .

Mercury exposure in Northeast North America

217

Fournier, F., Karasov,W.H ., Kenow, K.P., Meyer, M.W. and

Hines, R.K . (2002). The oral bioavailability and toxicoki-

netics of methylmercury in common loon

(Gavin irmner)

chicks.

Comp . Biochem . Physiol. Part A

133, 703-14 .

Frenzel, R.W. and Anthony, R.G. (1989). Relationship of diets

and environmental contaminants in wintering bald eagles

.

J. Wildl. Manage .

53, 792-802.

Furness, R.W. and Greenwood ., J .J .D. (1993)

.

Birds as mont-

tors

of

environmental change

.

New York, Chapman and

Hall .

Furness, R.W ., Lewis, S.A. and Mills, LA. (1990). Mercury

levels in the plumage of red-billed gulls

Larus novaehollan-

diae scopulinus

of known sex and

age . Environ. Pollut

.

63,

33-39

.

Furness, R.W ., Muirhead, S .J. and Woodburn, M. (1986).

Using bird feathers to measure mercury in the environment :

relationships between mercury content and moult

.

Mar.

Pollut

. Bull.

17, 27-30

.

Furulani, A. and Rudd, J.W.M. (1980). Measurement of mer-

cury methylation in lake water and sediment samples

.

Appl.

Environ. Microbial. 40,

770-76 .

Gardiner, E .E. (1972). Differences between ducks, pheasants,

and chickens in tissue mercury retention, depletion, and

tolerance to increasing levels of dietary mercury

.

Can. J.

Anim. Sri .

52, 419-23

.

Gardner, W.S., Kendall, D.R ., Odom, R.R., Windom, H.L

.

and Stephens, 1 .1. (1978). The distribution of methylmer-

cury in a contaminated salt marsh ecosystem

.

Environ .

Pollut

15, 243-51

.

Garibaldi, J.C., Jagoe, C.H. and Bryan, A.L. Jr. (1998). Dietary

exposure to mercury in nestling wood storks

(Mycleria

americana)

in Georgia .

Arch Environ. Contain. Toxicol

.

34,

398-405 .

Gerrard, P.M. and Louis, V .L.SI. (2001). The effects of exper-

imental reservoir creation on the bioacummulation of

methylmercury and reproductive success of tree swallows

(Tachycineta bicolor). Environ. Sci. Technal. 35,

1329-38 .

Gilmour, C.C., Henry, E.A. and Mitchell, R. (1992). Sulfate

stimulation and mercury methylation in freshwater sedi-

ments

.

Environ. Sci. Technol .'26,

2281-87.

Glass, G .E ., Sorensen, 1.A ., Schmidt, K.W. and Rapp, G.R .

Jr. (1990). New source identification of mercury contami-

nation in the Great Lakes

.

Environ

.

Sci.

Technol. 24,

1059-

69.

Gochfeld, M . (1980). Tissue distribution of mercury in normal

and abnormal young common terns

.

Marine Poll. Bull . 11,

362-77.

Gochfeld, M ., Belant,

1.E ., Shukla, T., Benson, T. and Burger,

J. (1996). Heavy metals in laughing gulls: Gender, age and

tissue differences .

Environ . Toxical. Chem.

15, 2275-83

.

Graveland, J. (1998). Effects of acid rain on bird populations

.

Environ.

Rev. 6, 41-54 .

Grier, J.W. (1974). Reproduction, organochlorines, and mer-

cury in northwestern Ontario bald eagles

.

Can. Field-Nat .

88, 469-75 .

Halbrook, R.S ., Brewer, R.L. Jr. and Buehler, D.A. (1999) .

Ecological risk assessment in a large river- reservoir : 7 .

Environmental contaminant accumulation and effects in

great blue heron.

Environ . Toxicol

Chem. 18, 641-48

.

---

218

Every el (it.

Hanes, R.S ., Rosenberg, K.V ., Lowe,

1.D ., Barker, S.E. and

Dhondt, A.A. (2002). Adverse effects of acid rain on the

distribution of the wood thrush

Hylocichla mustelina

in

North America

. Proc. Nat!

. Acad. Sci. 99, 11235-240.

Hammerschmidt, C.R. and Fitzgerald, W.F. (2004). Geo-

chemical controls on the production and distribution of

mcthylmercury in near-shore marine sediments

.

Environ.

Sci. Teehnol

. 38,1487-95 .

Hatch, W.R. and Ott, W.L. (1968). Determination of submi-

crogram quantities of mercury by atomic absorption spec-

trophotometry . Anal. Chem. 40, 2085-87

.

Heinz G. (1975). Effects of mcthylmercury on approach and

avoidance behavior of mallard ducklings

.

Bull. Environ .

Contain. Tnxicol. 13, 554.64 .

Heinz, G.11V and Hoffman, D J. (2003). Predicting mercury in

mallard ducklings from mercury in chorioallantoic mem-

branes . Bull. Environ

. Contam- Toxicol. 70, 1242-46 .

Heinz,GA. and Hoffman, D .J. (2004). Mercury accumulation

and loss in mallard eggs

. Environ

. Toxicol. Chem. 23, 222-24 .

Henny, C .1 .,

Blus,

L.J ., Grove, R.A and Thompson, S.P

.

(1991). Accumulation of trace elements and organochlorines

by surf scoters wintering in the Pacific Northwest . North-

wr .rt. Naturalist 72, 43-60

.

Henny, C .J .,

Hill, E.F ., Hoffman, D .J ., Spalding, M .G. and

Grove, R.A (2002). Nineteenth century mercury: Hazard to

wading birds and cormorants of the Carson River, Nevada

.

Ecotoxieology 11, 213--31

.

Hobson, K.A. (1990). Stable isotope analysis of marbled

murrelets: Evidence for freshwater feeding and determina-

tion of trophic level

. Condor 92, 897-903 .

Hohson, K . A. and Scaly, S. (1991). Marine protein contribu-

tions to the diet of Northern Saw-whet Owls on the Queen

Charlotte Islands : A stable isotope approach

.Auk 108, 437-

40 .

Hobson, K.A., Sirois, J. and Gloutney, ML . (2000). Tracing

nutrient allocation to reproduction with stable isotopes: A

preliminary investigation using colonial waterbirds of Great

Slave Lake

. Auk 117, 760-74

.

Hobson, K.A. (1987). Use of stable-carbon isotope analysis to

estimate marine and terrestrial protein content in gull diets

.

Can. J. Zool . 65, 1210-13 .

Hohson, K.A. (1993). Trphic relationships among high Arctic

seabirds: Insights from tissue-dependent stable-isotope

models

. Mar_ Ecol. Prog

. Sec 95, 7-18

.

Hobson, K.A ., Hughes, K.D, and Ewins, P .J. (1997). Using

stable-isotope analysis to identify endogenous and exoge-

nous sources of nutrients in eggs of migratory birds :

Applications to Great Lakes contaminants research

. Auk

114,467-78 .

Hoffman, R.D. and Curnow, R.D. (1979). Mercury in herons,

egrets and their foods. J. Wild(. Manage. 43, 85-93 .

Jackson, T.A. (1988)_ The mercury problem in recently formed

reservoirs of northern Manitoba (Canada): Effects of

impoundment and other factors on the production of

mcthylmercury by tnicoorganisns in sediments

. Can. J.

Fish . Aquat. Sci. 45, 97 121

.

Kambamandi-Dim ou, A., Kamarianos, A. and Kilikidis. S

.

(1991). Transfer of mcthylmercury to hens' eggs after oral

administration

. Bull. En 'iron-Coolant Tnxieol .46, 128-33

.

Kamman, N. and Engstrom, D. (2002). Historical and present

fluxes of mercury to Vermont and New Hampshire lakes

inferred from 21 OPh dated sediment cores . Alm

. Environ

. 36,

1599-1609 .

Kamman, N.C ., Burgess, N.M ., Driscoll, C.T., Simonin, H.A .,

Goodale, W . M., Linchan, J., Estahrook, R ., Hutcheson,

M., Major, A., Scheuhammer, A.M. and Sermon, D.A

.

(2005). Mercury in freshwater fish of northeast North

America - :t geographic perspective based on fish tissue

monitoring databases . Eemovicology 14, 163-80 .

Kenow, K.P .,

Gutreuter, S., Hines, R.K .,

Meyer, M.W.,

Fournier, FF and Karasov, W.H. (2003). Effects of methyl

mercury exposure on the growth ofjuvenile common loons

.

Ecotoxicologr 12, 171-82

.

Ken.. .

K .P .,

Gutrcutcr, S .. Hines,

R.K .,

Meyer, M.W .,

Foamier, F- and Karasov, W.H . (2003). Effects of methyl

mercury exposure on the growth ofjuvenile common loons

.

Eaotoxieologv 12, 171-82

.

Kidd, K . A. (1998). Use of stable isotope ratios in freshwater

and marine biotnagnification studies . In J. Rose (eds).

Environmental toxicology Current developments., pp- 357-

376. Amsterdam: Gordon Breach Science

.

Knight, R . L., Randolph, P .) ., Allen. G . T., Young, L.S. and

Wigen, R .J. (1990) Diets of nesting bald eagles, Haliaeetms

leucocepludus, in western Washington. Can . Field-Nat . 104,

545-51

.

Kozic, KU. and Anderson, R . K. (1991). Productivity, diet, and

environmental contaminants in bald eagles nesting near the

Wisconsin shorelinc of take Superior

. Arch . Environ . Con-

tern- To.dcof 20, 41-48 .

Kramer, D-, Goodale, W., Kennedy,W.L., Carstensen, B, and

Kaur, 1'. (2005). Relating land cover characteristics and

common loon mercury levels using Geographic Information

Systems . Ecrnoxieologv 14, 253-62.

Lamhorg, C.H ., Fitzgerald, W.F., Skoog, A. and Visscher, P.T

.

(2005). The abundance and source of mercury-binding

organic ligands in Long Island Sound

. Marine Chem. (in

press) .

Lane, OP., Evers, D.C ., Alhano, D., Haines, T. and Taylor, R .

(2004). Belted kingfishers

(Ceryle alcyon) as indicators of

mcthylmercury availability in aquatic systems. Unpubl. re-

port BRI 2004-13 submitted to the Maine Department of

Environmental Protection- BioDiversity Res . Inst., Gor-

ham, Maine

.

Langis, R., Langlois, C, and Morneau, F. (1999). Mercury in

birds and mammals. In M. Lucott, R. Schetagne, N. The-

rien, C. Langlois and A- Tremblay (eds)

. Mercury in the

biogeoehemiral cycle ., pp. 131-144. New York: Springer

.

Lobring, L.B& and Potter, B .B. (1991). Method 245.6. Rev. 2 .3

Determination

of

mercury in tissues

by cold vapor atomic

absorption

.spectrometry

.

USEPA/600/4-91/010. Methods

for the determination of metals in environmental samples,

pp. 281-293. Washington, D.C

.

Lucoue, M ., Schetagne, R., Thericn, N„ Langlois, C. and

Tremblay, A., (eds). (1999)- Mercury

in the hiogeodtemical

c-ele- New York, Springer-

Mason, R., Abbot, M., Bodaly, D., Bullock, R ., Driscoll, C .

.

Evers, D., Lindherg, S., Murray, M . and Swain, E. (2005)

.

Monitoring the environmental response to changes in mer-

---

cury contamination from the atmosphere: A multi-media

challenge .

Environ

. Sci. Technol .

39: 15A-25A

.

Meyer, M.W., Evers, D.C., Daulton, T. and Braselton, W.E

.

(1995). Common loons

(Gavia immer)

nesting on low pH

lakes in northern Wisconsin have elevated blood mercury

content .

Water, Air, Sail Pollut

80, 871-80

.

Meyer, M .W ., Evers, D.C ., Hartigan, J .J. and Rasmussen, P .S .

(1998). Patterns of common loon

(Gavia immer)

mercury

exposure, reproduction, and survival in Wisconsin, USA

.

Environ . Toxicol

.

Chem. 17, 184-90 .

Miller, E.K ., VanArsdale, A ., Keeler,

1 .G., Chalmers, A .,

Puissant, L., Kamman, N. and Brulotte, R. (2005). Esti-

mation and mapping of wet and dry mercury deposition

across northeastern North America

.

Ecatoxicology 14 .

Mizutani, H., Fukuda, M., Kabaya, Y. and Wads, E . (1990) .

Carbon isotope ratio of feathers reveals feeding behavior of

cormorants

.

Auk

107, 400

.37

.

Monteiro, L.R. and Furness, R. W. (1995). Seabirds as monitors

of mercury in the marine environment,

Water Air Soil

Polhrt .

80, 851-70 .

Monteiro,

L.R. and Furness, R. W. (1997). Accelerated increase

in mercury contamination in North Atlantic mesopelagic

food chains as indicated by time series of seabird feathers

.

Environ . Toxicol. Chem.

16,2489-93

.

Monteiro, L.R. and Furness, R.W. (2001). Kinetics, dose-

response, excretion, and toxicity of methylmercury in free-

living Cory's shearwater chicks

.

Environ. Toxicol. Chem .

20,

1816-24.

Moore, D.R .J

., Sample,B.E ., Suter,G.W ., Parkhurst, B.R. and

Teed, R.S. (1999). A probabilistic risk assessment of the

effects of methylmercury and PCBs on mink and kingfishers

along East Fork Poplar Creek, Oak Ridge, Tennessee, USA

.

Environ . Toxicol.

Chem. 18, 2941-53 .

Morel, F.M ., Kraepiel, A.M. and Amyot, M. (1998). The

chemical

cycle

and bioaccumulation of mercury

.

Annu. Rev.

Ecol . Syst

29, 543-66.

Nichols, J. and Bradbury, S. (1999). Derivation of wildlife

values for mercury . J .

Toxicol

. Environ. Health, Part B, 2,

325-55 .

Nisbet, I.C.T., Montoya J, P., Burger, ). and Hatch, J.J. (2002)

.

Use of stable isotopes to investigate individual differences in

diets and mercury exposure among common terns

Sterna

hirundo in

breeding and wintering grounds .

Marine Enrol.

Progress Ser

.

242, 267-74

.

Ohlendorf, H .M

., Lowe, R.W., Kelly, P.R. and Harvey, T.E .

(1986). Selenium and heavy metals in San Francisco Bay

diving ducks .

J. Wild!. Manage .

50, 64-71

.

Pearce, P.A., Price, 1 .M . and Reynolds, L.M . (1976). Mercury

in waterfowl from eastern Canada . J

.

Wildl. Manage .

40,

694-703 .

Piper, W ., Paints,

1.D ., Evers, D.C., Meyer, M.W

., Tishler,

K.B ., Klich, M. and Hartigan, J .J . (1997) . Local movements

of color-marked common loons

.

J.

Wild!. Manage .

61,

1253-61 .

Pokras, M.A ., Hanley, C. and Gordon, Z. (1998). Liver mer-

cury and methylmercury concentrations in New England

common loons

(Gavin immer) . Environ . Toxicol. Chem. 17,

202-04 .

Mercury exposure in Northeast North America 219

Pokras, M.A ., Rohrbach, S., Press, C., Chard, R., Perry, C.

and Burger, J. (1992) .

Environmental pathology of 124

common loons from the northeastern United States 1989-

1992The loan and its ecosystem: Status, management, and

environmental concerns, pp .

20-33. Concord, New Hamp-

shire: U.S. Fish Wildl. Serv .

Porvari, P., Verta, M., Munthe, J. and Haapancn, M . (2003).

Forestry practices increase mercury and methyl mercury

output from boreal forest catchments

.

En viron. Sci. Technol.

37,2389-93 .

Reynolds, K.D .,

Rainwater, T.R., Scollon, E.1 ., Sathe, S.S .,

Adair, B.M. and Dixon, K.R. (2001). Accumulation of

DDT and mercury in prothonotary warblers

(Protonomrio

citrea)

foraging in a heterogeneously contaminated arvi-

ronment.

Environ. Toxicol . Chem .

12, 2903-09.

Rimmer, C.C., McFarland, K.P ., Evers, D.C., Miller, EX

.

Aubry, Y., Busby, D. and Taylor, R

.J . (2005). Mercury

levels in Bicknell's thrush and other insectivorous

passerine

birds in montane forests of the

northeastern

United States

and Canada .

Ecatoxicology

14: 223-40 .

Rumbold, D.G ., Niemczyk, S.L., Fink, L.E., Chandrasckhar,

T., Harkanson, B. and Lane, K.A. (2001). Mercury in

eggs and feathers of great egrets

(Arden alhus)

from the

Florida Everglades

.

Arch .

Environ

.

Contam

.

Toxicol. 41,

501-07 .

Louis, V.L.St., Rudd, JW.M., Kelly, C.A ., Hall, R.D ., Roflus,

K.R., Scott, K .J ., Lindberg, S.E. and Dong, W. (2001).

Importance of the forest canopy to fluxes of methyl mercury

and total mercury to boreal ecosystems .

Environ . Si.

Technol.35,3089-98 .

SAS Institute, Inc. (2001) .

JMP Version 4.01,

Cary, North

Carolina, SAS Institute Inc.

.

Scheuhammer, A.M. (1987). The chronic toxicity of aluminum,

cadmium, mercury, and lead in birds : A review .

Environ.

Pa1lut .

46,263-95 .

Scheuhammer, A.M. (1988). Chronic dietary toxicity of meth-

ylmercury in the zebra finch,

Poephila guffaw. Bull. Environ

.

Conwm. Toxicol .

40, 123-30.

Scheuhammer, A.M. and Blancher, P.) . (1994)

.

Potential

risk

to common loons

(Gavia immer)

from methylmercury

exposure in acidified lakes

.

Mydrohiologia

279/280, 445-

55 .

Scheuhammer, A.M ., Perrault, 1A. and Bond, D . E. (2001)

Mercury, methylmercury, and selenium concentrations in

eggs of common loons

(Gavia immer)

from Canada.

Envi-

ron. Moral. Assess

.

72, 79-84 .

Scheuhammer, A.M ., Wong, A.H.KK and Bond, D. (1998a) .

Mercury and selenium accumulation in common loons

(Gavia

immer) and common mergansers

(Mergus mergan-

ser)

from eastern Canada

.

Environ. Toxicol.

Chem. 17,

197-201

.

Scheuhatmncr, A.M .,

Atchison,

C.M ., Wong, A .H .K. and

Evers, D.C. (1998b). Mercury exposure in breeding com-

mon loons (Gavin

immer)

in central Ontario, Canada .

Environ . Toxicol.

Chem. 17, 191-96

.

Schmutz, 1.A. and Hobson, K .A. (1998). Geographic, tempo-

ral, and age-specific variation

in

diets of glaucous gulls

in

western Alaska

.

Condor 100,

119-30

.

---

220

Evers et a! .

Schuster, P.F., Krabbenhoft, D.P., Nafiz, D- L., Cecil, L.D .,

Olson, M.L ., Dewild, 1.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 .

Environ . Sci. Tech.

36,

2303-10 .

Seilcr, R.L., Lico, M .S., Wiemeycr, S.N. and Even, D.C .

(2004). Mercury in the Walker River Basin, Nevada and

California-Sources, distribution, and potential effects on

the ecosystem

.

U.S. Geo. Sure-, Sri. Invest . Rept. 2004-5147 .

Shriver, W . G., Evers, D.C. and Hodgman, T.P. (2002). Mer-

cury exposure profile for Sharp - tailed Sparrows breeding

in coastal Maine salt marshes . Unpuhl . report BRI 2002-I I

submitted to the Maine Dept. Environ- Protection

. BioDi-

versity Res. Inst., Gorham, Maine

.

Spalding, M.G ., Frederick, P.C: ., McGill, H.C., Bouton, S.N.

and McDowell, L .R. (2000a). Methylmercury accumulation

in tissues and its effects on growth and appetite in captive

great egrets.

J. Wild! .

Di.v. 36, 411-22.

Spalding, M.G ., Frederick, P.C ., McGill, H.C ., Bouton, S.N .,

Richey, L .1 ., Schumacher, LM., Blackmore, C-QM. and

Harrison, 1. (2000h). Histological, neurologic, and imuno-

logic effects of methylmercury

in captive great egrets .

J

.

Wild!- Dis.

36, 423-35

.

Stoewsand, G . S., Bache, C .AA

and Lists, D .3. (1974). Dietary

selenium protection of methylmercury intoxication of Jap-

anese quail

.

Bull. Environ . Contam. Toxicol.

11, 152-56.

Sundlof, S . F., Spalding, M.G ., Wentworth, J.D, and Steible,

C.K. (1994). Mercury in livers of wading

birds (Ciconii-

formes) in southern Florida-

Arch. Environ. Contara . Taxi-

col.

27, 299-305 .

Thompson, D.R (1996)-

Mercury in birds and terrestrial

mammals. In W.H. Beyer, G.H-Hcinz and A .W. Redmond-

Norwood (eds)

.

Environm ntal contaminants

in wildlife :

Interpreting !issue concentrations, pp-

341-356. Boca Raton,

Florida: Lewis Publishers

.

Thompson, D.R. and Furness, R . W. (1989)

. Comparison of

total and organic mercury levels in seabird feathers

.

Marine

Pollut. Bull.

20, 577-79.

Thompson, D .R ., Hamer, K.C. and Furness, R.W. (1991)

Mercury accumulation in great skuas

Catharacia skua

of

known age and sex, and its effect upon breeding and

sur-

vival .

J. App!. Ecol

.

28, 672-84 .

Thompson, D.R ., Stewart, F.M. and Furncss, R.W. (1990).

Using seabirds to monitor mercury in marine environments

:

The validity of conversation ratios for tissue comparisons

.

Marine Pollut . Bull

.

21, 339-42 .

United Nations Environment Programme- (2003)

.

Global mer-

cury assessment.

Geneva, Switzerland, UNEP Chemicals

.

U .S. EPA. (1997) .

Mercury study sport to Congress.

EPA-452/

R-97-008. U .S Environ . Protection Agency . Washington,

D.C .

U .S. EPA. (1998) .

Mercury in solid% and solutions by thermal

decomposition, amalgamation, and atrnnic absorption spec-

trophotometry.

Method 7473- U.S

.

Environ. Protection

Agency, Washington, DC-

US. EPA. (2002). Workshop on the fate, transport, and trans-

formation ofmercury in aqua tic and terrestrial environments

.

EPA/625/R-02/005. U .S. Environ. Protection Agency, Cin-

cinnati, OH and U-S . Geological Survey, Reston, Virginia

.

U.S. Fish and Wild life Service (2003).

Evaluation ofthe Clean

Water Act Section

304(a)

human health

criterion for meth-

ylmercury :

Protectiveness for threatened

and endangered

wildlife

in California .

Sacramento, California, U .S .

Fish

Wild[. Serv

.

.

VanArsdale, A ., Weiss, 1., Keeler, G .J ., Miller, E.K ., Boulet,

G., Brulotte, R. and Puissant, L. (2005). Patterns of mer-

cury deposition in northeastern North America

.

Ecotoxi-

cology

14, 37-52 .

Wayland, M., Gilchrist, H.G ., Marchant, T., Keating, J. and

Starts, J.E. (2002)- Immune function, stress

response, and

body condition in arctic-breeding common eiders in relation

to cadmium, mercury, and selenium concentrations

.

Envi-

ron.

Research

Section A

90, 47-60 .

Weech, S . A., Scheuhammer, A .M., Elliot, 1.E. and Cheng,

K.M. (2004)- Mercury in fish from the Pinchi

Lake re-

gion, British Columbia,

Canada .

Environ

.

Pulhu .

131,

275-86 .

Welch, L. (1994)

.

Contaminant hardens and reproductive rates of

bald eagles breeding in Maine .

M.S. thesis, Univ. Maine,

Orono, Main,

.

Wiemeycr, S.N ., Lamont, T.G., Bunck, C.M ., Sindelar, C.R .,

Gramlich, F.J

., Fraser, J. D. and Byrd, M.A. (1984). Orga-

nochlorine pesticide, polychlorobiphenyl, and mercury res-

idues in

bald eagle eggs -

1969-1979

-

and their

relationships to shell

thinning and reproduction

.

Arch

.

Environ. Contain. Toxicol .

13, 529-49.

Wiener, J.G. and Shields, P .1. (2000). Mercury in the Sudbury

River (Massachusetts U.S.A .) : Pollution history and syn-

thesis ofrecent research

.

Can . J . Fish

.

Ayuat.Sci.57,1053-61 .

Wiener, 1G. and Spry- D]- (1996) . Toxicological significance

of mercury in freshwater fish . In W.N. Beyer, G.H

. Heinz

and A W. Redmon-Norwood (eds)

.

Environmental contam-

inants fn mild/ii/cc Interpreting tissue concentrations,

pp. 297-

339. Boca Raton, Florida: Lewis Publishers

.

Wiener, J . G., Martini, R.E .,

Shelly, T.B. and Glass, G.E

.

(1990). Factors influencing mercury concentrations

in

walleyes in northern Wisconsin lakes

.

Trans. Am .

Fish. So,

.

119, 862-70

.

Wiener,

1.G.,

Krabbenhoft, D. P., Heinz, G.H. and Scheu-

hammer, A.M. (2003) .

Ecotoxicology of mercury. In D .J .

Huffman, B.A. Rattner, G.A. Burton Jr. and J. Cairns Jr.

(ads). Handbook of ecotaxieology .,

pp. 409-463. Boca Ra-

ton, Florida: Lewis Publishers

.

Winfrey, M.RR and Rudd, J .W.M . (1990). Environmental

five-

ban

affecting the formation of methylmercury in low pH

lakes .

Environ. Toxicol. Chem

. 9,

853-69 .

Wolfe, M .F . and Norman, D . (1998)

.

Effects of waterborne

mercury on terrestrial wildlife at Clear Lake : Evaluation

and testing of a predictive model .

Environ. Toxicol. Chem

.

17,214-27.

Wolfe, M.F ., Schwarzhaeh, S-and Sulaiman, R-A . (1998). Ef-

fects of mercury on wildlife :

A comprehensive review

.

Environ. Toxicol . Chem .

17, 146-60 .

Xun, L., Campbell, N .E.R. and Rudd, J.W.M

. (1987). Mea-

surement of specific rates of net methy[mercztry production

---

Mercury exposure in Northeast North America

221

in the water column and surface sediments of acidified and

Zar, J.H. (1999) .

Rio.statistical analysis

.

Prentice Hall, tipper

circumneutral lakes

.

Can. J. Fish. Aquat. Sci. 44, 75(1-57

.

Saddle River, New Jersey

.

---

Arch. Environ. Contam. Toxicol. 44, 257-264 (2003)

DOI :

10.1007/s00244-002-2021-6

Embryotoxic Thresholds of Mercury : Estimates from Individual Mallard Eggs

G. H. Heinz, D. J. Hoffman

Patuxent Wildlife Research Center, U .S. Geological Survey, 11510 American Holly Drive, Laurel, Maryland 20708-0017,

USA

Received: 9 February 2002/Accepted: 12 June 2002

Abstract. Eighty pairs of mallards (Anas platyrhynchos) were

fed an uncontaminated diet until each female had laid 15 eggs .

After each female had laid her 15th egg, the pair was randomly

assigned to a control diet or diets containing 5, 10, or 20 µg/g

mercury as methylmercury until she had laid a second set of 15

eggs. There were 20 pairs in each group. After the second set

of 15 eggs, the pair was returned to an uncontaminated diet,

and the female was permitted to lay another 30 eggs . For those

pairs fed the mercury diets, the even-numbered eggs were

incubated and the odd-numbered eggs were saved for possible

mercury analysis. Mercury in the even-numbered eggs was

estimated as the average of what was in the neighboring odd-

numbered eggs. Neurological signs of methylmercury poison-

ing were observed in ducklings that hatched from eggs con-

taining as little as 2.3 pg/g estimated mercury on a wet-weight

basis, and deformities were seen in embryos from eggs con-

taining about I p

.g/g estimated mercury . Although embryo

mortality was seen in eggs estimated to contain as little as 0 .74

p

.g/g mercury, there were considerable differences in the sen-

sitivity of mallard embryos, especially from different parents,

with some embryos surviving as much as 30 or more p.g/g

mercury in the egg .

Mercury contamination of eggs has been suspected as a cause

of impaired reproduction in wild birds (Barr 1986 ; Fimreite

1974), but associations between mercury levels in eggs and

reproductive problems in field studies are complicated by the

presence of other environmental stressors that may have an

effect. Consequently, the results from controlled laboratory

studies with mallards, black ducks

(Anus rubripes),

ring-

necked pheasants (Phasianus colchicus), and chickens (Gal-

lus gallus) have been used to help establish the concentrations

of mercury in eggs that are associated with effects on avian

reproduction (Heinz 1974, 1976, 1979; Finley and Stendell

1978; Fimreite 1971 ; Tejning 1967). The laboratory data have

been used as a guideline for whether the mercury concentra-

tions found in the eggs of wild birds might be harmful (Eisler

2000; Henny el al . 2000; Meyer el al. 1998; Scheuhammer

el

Correspondence to: G.

H.

Heinz; email' gary

heinz@usgs.gov

al. 2001; Thompson 1996; Wiemeyer el al. 1984; Wolfe el al

.

1998) .

In past laboratory studies with mallards, groups of breeding

adults were fed different concentrations of methylmercury, and

a mean mercury level was calculated from a sample of eggs

from each group (Heinz 1974, 1976, 1979; Heinz and Hoffman

1998). When reproductive success of a group was significantly

less than that of controls, the mean mercury level in eggs from

that group was judged to be a harmful level . However, indi-

vidual embryos likely differ in their sensitivity to methylmer-

cury. As long as laboratory-generated findings with methyl-

mercury continue to be used to help protect the reproductive

success of wild birds, especially species that cannot tolerate

much reproductive loss because their numbers are already low,

it is important to know how sensitive individual embryos can

be .

Our primary objective was to estimate the lowest concentra-

tions of mercury in mallard eggs that would harm the most

sensitive embryos. Our secondary objective was to determine

how variable individual mallard embryos are in their sensitivity

to methylmercury .

Materials and Methods

Care of Adults and Administration of Mercury Diets

Eighty breeding pairs of I-year-old mallards

(Anas platyrhynchos)

(Kidder Game Farm, Milton, WI) were randomized to I m 2 outdoor

breeding pens on April 15, where they had access to flowing water and

a commercial game bird breeder diet (Purina Mills, St. Louis, MO)

.

The diet contained about 20% crude protein, 2 .5% crude fat, and 7%

crude fiber. On April 24 we began collecting eggs from each pair . Eggs

were labeled, stored in a Kuhl egg cooler (Flemington, NJ) at about

13-14°C, and at weekly intervals incubated in a Kuhl incubator at a

temperature of 37 .5°C and a relative humidity of about 54% . On day

25 of incubation we transferred the eggs to a Kuhl hatcher set at a

temperature of 37 .2°C and a relative humidity of about 70%

.

Each of the 80 pairs was fed the untreated commercial diet until the

female had laid 15 eggs. These first 15 eggs served to establish

baseline data on the hatching success of each pairs' eggs prior to the

start of mercury treatment. On the day a female laid her 15th egg, that

pair was randomly assigned to one of the four treated diets : a control

diet or diets containing 5, 10, or 20 p

.g/g mercury as methyhnercury

chloride. Methylmercury has been shown to be the main form of

A R H I V E S O F

Environmental

Contamination

a n d

Toxicology

0 2003

springer-Verlag New York

Inc.

---

258

mercury found

in the eggs of wild birds (Rumbold

et at

2001

;

Scheuhammer et a! . 2001). Each treatment had 20 pairs of ducks. The

dietary concentrations of mercury we used, especially the 10 and 20

p-g/g treatments, were higher than had been used in some previous

studies with mallards (Heinz 1974, 1976, 1979), but these concentra-

tions were selected because we wanted mercury levels in eggs to

increase greatly over a short period of time

.

The methylmercury diets were prepared by first dissolving methyl-

mercury in a small amount of acetone and then in a larger volume of

corn oil. These solutions were then mixed with the breeder diet to

make premixes, which were kept frozen until needed . The premixes

were blended into a larger volume of breeder diet to make the final

diets; these final diets contained about 0.2% corn oil. Two samples of

each diet were analyzed to confirm the concentration of mercury . The

samples of control diet were reported to contain less than 0 .02 µg/g

mercury. Recovery of mercury from the three treated diets averaged

89%. Mercury was analyzed .by cold-vapor atomic absorption spec-

trophotometry at the Patuxent Analytical Control Facility, located at

the Patuxent Wildlife Research Center .

For the 20 pairs selected to receive each of the three mercury-treated

diets, only the first 14 eggs were incubated to establish a baseline for

hatching success; the 15th egg was saved for possible mercury anal-

ysis. For the 20 pairs of controls, all of the first 15 eggs, plus all later

eggs, were incubated . Eggs 16 through 30 were collected while the

female was being fed her treated diet. On the day the 30th egg was laid,

the pair was switched back to an untreated diet, and the female was

allowed to lay another 30 eggs or until July 16, when the study was

terminated. For those females switched to one of the three mercury-

containing diets, we only incubated their even-numbered eggs and

saved their odd-numbered eggs for possible mercury analysis .

Determining the concentrations of mercury that harmed individual

eggs required some way of estimating how much mercury was in an

egg while still allowing that egg to go through all the stages of

incubation, hatching, and duckling survival. Our approach to estimat-

ing the mercury concentration in an incubated, even-numbered egg

was to take the average of the mercury in the odd-numbered egg laid

the day before and the odd-numbered egg laid the day after the

even-numbered egg. During the period when an egg-laying female was

on a constant methylmercury diet, mercury levels in eggs would be

increasing; once the mercury diet was stopped mercury in eggs would

decrease. In either case, the average of the mercury between the

previous egg and subsequent egg should give a good estimate of what

was in the unanalyzed egg

.

We examined the fate of all the control eggs and all the even-

numbered eggs of mercury-treated pairs . After eggs hatched, the

ducklings were banded and kept for 6 days in heated pens provided

with flowing water and untreated duck starter diet (Purina Mills)

.

Because many hundreds of odd-numbered eggs were saved for poten-

tial mercury analysis, it was not feasible to have them all analyzed for

mercury. Therefore, after the study of the hatching success of the

even-numbered eggs and the 6-day survival of ducklings was com-

pleted, we carefully examined the fate of the eggs from each of the

mercury-treated females . Some of the pairs had poor hatching success,

even prior to being put on their mercury diets. We did not analyze

many odd-numbered eggs from these poor-achieving pairs because

failure of an egg to hatch could just as easily have been due to chance

as to mercury poisoning. We limited most of our mercury analyses of

odd-numbered eggs to situations where an even numbered egg (I)

produced a deformed embryo, (2) produced a duckling that displayed

neurological signs of methylmercury poisoning, or (3) failed to hatch

when a series of eggs laid before it did hatch . In addition, with a few

exceptions that are noted in the Results and Discussion, we did not

include estimates of mercury in even-numbered eggs when more than

I day

separated the laying of the even-numbered egg from the laying

of the neighboring odd-numbered eggs. Our estimates of mercury in

these selected, even-numbered eggs were most likely to provide in-

formation related to true mercury toxicity, as contrasted to the normal

G . H. Heinz and D. J. Hoffman

amount of embryo mortality that even controls might be expected to

suffer during incubation. Even with these restrictions, we analyzed

over 200 odd-numbered eggs for mercury and related the estimates of

mercury in even-numbered eggs to the toxic effects listed later .

Neurological Signs

The feeding of methylmercury to breeding mallards and black ducks

has been shown to produce neurological signs related to characteristic

brain lesions in some of their ducklings (Heinz and Locke 1976 ; Finley

and Stendell 1978). Affected ducklings may appear normal when

hatched, but within I to 3 days they begin to lose coordination and

stagger about. In several studies, including the current one, we have

never seen a control mallard duckling exhibit these neurological signs

.

Therefore, when ducklings from mercury-treated parents in the current

study began to lose balance and stagger, this was an almost certain sign

that they had been poisoned by methylmercury

.

Deformities

Methylmercury has been reported to cause various kinds of deformi-

ties in mallard embryos, but deformities also occur in a small percent-

age of control ducklings (Heinz and Hoffman 1998; Hoffman and

Moore 1979). In one large study with mallards, 6.1% of 1-week or

older control embryos exhibited some form of deformity, primarily

hydrocephaly (Heinz and Hoffman 1998) . In the same study, 16.4% of

the embryos from parents fed 10 p.g/g mercury as methylmercury were

deformed, many exhibiting deformities not seen in controls . In the

current study, we could not be as certain that deformed embryos were

caused by mercury as we were that neurological signs were the result

of mercury in the egg .

Failure of Eggs to Hatch

Not all eggs, even controls, can be expected to hatch. A hatch of

70-80% of fertile control eggs is a good hatch for mallards (Gile and

Meyers 1986; Heinz 1979; Heinz and Fitzgerald 1993; Heinz and

Hoffman 1998). Therefore, to reduce the likelihood that the failure of

an egg to hatch was unrelated to mercury treatment, we included data

only for those pens that had a proven record of good hatching success

prior to the switch to the mercury diets. After subtracting out the one

or two unusable eggs that were occasionally cracked or infertile among

the first 14 laid by the mercury-treated females, we defined good

hatching success as there being no more than I unhatched egg out of

the first 12, 13, or 14 usable eggs . Although not as certain an indicator

of methylmercury poisoning as the exhibiting of neurological signs,

the failure of an egg to hatch (when other uncontaminated or less

contaminated eggs laid before or after this egg did hatch) provided

information on threshold levels of mercury in mallard eggs that cause

harm.

Mercury in Eggs That Hatched and from Which the

Hatchlings Survived 6 Days

Although the main purpose of our study was to determine the lowest

concentrations of mercury that harmed the most sensitive mallard

embryos, we also were interested in how much embryos could differ

in their sensitivity to mercury. One way to examine this variability was

to determine how much mercury could be in an egg that did hatch and

the duckling survived for the 6-day observation period. When an

---

Mercury Thresholds in Eggs

embryo died or was deformed, we could not be absolutely certain that

mercury was to blame. In contrast, we did know with absolute cer-

tainty when the estimated level of mercury in an egg did not

kill an

embryo, cause a deformity, or cause neurological signs

in the hatch-

ling. By examining the concentrations of mercury in eggs

that pro-

duced healthy 6-day-old ducklings, we were able to understand how

different mallard embryos can be in their sensitivity to methylmercury

.

Results and Discussion

Neurological Signs

The lowest estimated mercury concentration

in an egg that

produced a duckling with neurological signs was 2.3 pg/g on a

wet-weight basis and the highest value was 30 p.g/g (Table 1) .

Given the specificity of this sign as something seen in previous

studies only in mercury-poisoned ducklings, we believe the

ducklings listed in Table I were poisoned by mercury and the

estimated mercury levels in eggs produced these neurological

problems. There were no control ducklings in this study

that

ever exhibited a loss of balance and staggering

. Also, in the

three groups that were switched to mercury-treated diets, there

were no ducklings that exhibited neurological signs when they

hatched from eggs laid prior to the start of the mercury diets .

Most of the ducklings that exhibited signs of mercury poison-

ing hatched from eggs 22 through 30, which was during the

period when mercury was in the diet of the mothers. Two

ducklings that exhibited neurological signs hatched from eggs

that were laid shortly after the cessation of the mercury diet

.

Deformities

The lowest estimated mercury concentration associated with a

deformed embryo was 0.93 µ.g/g on a wet-weight basis and the

highest value was 18 µg/g (Table 2). The estimate of 0.93 pg/g

for egg number 62-48 (the 48th egg laid by female number 62)

was based on the average between a value of 0.935 for egg 47

and 0.926 for egg 49. Egg 62-46 also produced a deformity and

contained only 1 .0 pg/g mercury, which was the average

between values of 1 .15 pg/g for egg 62-45 and 0.935 pg/g for

egg 62-47. These two estimates seem to be very accurate

averages based on mercury in the neighboring odd-numbered

eggs. Female 62 also produced deformed embryos in her 20th

egg, which contained 3.7 pg/g mercury, and in her 54th egg .

The 54th egg is not listed in Table 2 because the odd-numbered

neighboring eggs were not analyzed for mercury, but we know

that the 54th egg almost certainly contained less mercury than

the value of 0.926 pg/g reported for egg 49. The embryo from

this 54th egg was badly deformed, with short legs, extra toes,

a spoon-shaped upper bill, and an abnormal right wing . An

unusual pattern of embryotoxic responses was seen in the eggs

from female 62 in that, after egg number 20, which had a

deformed embryo, eggs 22, 24, 26, and 30 (which would have

contained more mercury than 22) all hatched . Egg number 28

was infertile. All of the even-numbered eggs from 32 through

44 also hatched, but they would have had more mercury than

egg numbers 46, 48, and 54, which produced deformed em-

Table l. Estimated mercury concentration in eggs that produced

ducklings that exhibited neurological signs of methylmercury poi-

soning

259

:, The pen number of the female that laid the egg, followed by the egg

number (according to the sequence in which the eggs were laid)

.

b The concentration of mercury in the egg producing this duckling was

estimated by taking the average of the mercury in the egg laid the day

before and the egg laid the day after this egg.

bryos. We do not know why, but the embryos from this female

seemed to be very different in their sensitivity to mercury

.

Another female's egg (65-16) also produced

a deformed

embryo and contained an estimate of only 1 .0 pg/g mercury .

The 15th

egg by

female 65 was laid the day before the mercury

treatment was started and was reported to contain 0 .0278 pg/g

mercury. The 17th egg was reported to contain

2.06 pg/g

mercury. The estimate, based on eggs

15 and 17, could be

biased toward the low side because the 15th egg was

a pre-

treatment egg, and mercury could have jumped considerably by

the 16th egg. Like female 62, female 65 also had an unusual

pattern of embryotoxic effects in her eggs, suggesting differ-

ences in sensitivity among siblings . After egg 16, eggs

18, 20,

and 22 all hatched. Thereafter, all of her eggs failed to hatch,

either because they died (in two cases) or were infertile

in all

the rest.

In addition to the deformities listed in Table 2 or discussed

so far, there were four additional embryos

from mercury-

treated parents that exhibited deformities, but either the neigh-

boring eggs were not analyzed or the neighboring eggs were

not laid within I day of the egg that produced the deformed

embryo. One of these deformities, in the 32nd egg laid by a

female fed 5 pg/g mercury was a deformed ball of tissue

without any limbs. Only the 33rd egg from this

female was

analyzed and it contained 1 .1 pg/g mercury; therefore, the

deformed 32nd egg probably contained somewhat more than

1 .1 pg/g mercury because it was laid one day closer to the time

when the female was still being fed mercury

.

A second additional embryo, from the 30th egg laid by

a

female fed 5

tg/g mercury, had a clubbed right foot

that

Concentration of

Mercury in Parents'

Diet (pg/g)

Female's Pen

Number-Egg

Number'

Mercury Estimated to

Have Been in Egg

(µg/g, wet weight)^

10

64-24

10

20

2-24

2.7

20

2-26

4.0

20

6-24

22

20

6-28

22

20

6-30

26

20

9-26

27

20

9-34

24

20

20-24

3 .6

20

20-26

3 .3

20

27-26

5 .4'

20

28-22

2.3

20

45-26

4 .1

20

46-30

8 .1

20

46-32

7.8

20

52-28

30

20

63-22

19

20

77-24

23

---

260

Table 2. Estimated mercury concentration in eggs that produced embryos with deformities

pointed backward plus rigid joints in its leg. It came from an

egg that was estimated to contain 8 .7 ag/g mercury, but the

estimate was based on the average mercury in the 29th egg

(10.1 log/g), which was laid 2 days prior to the egg with the

deformed embryo, and the 31st egg (7 .29 p.g/g), which was laid

I day after. If anything, the 30th egg may have had slightly

more mercury than the estimate.

A third embryo, from the 32nd egg laid by a female fed 10

p

.g/g mercury, had two heads partially fused together, a twisted

upper bill, and a deformed right wing . Only the 33rd egg was

analyzed for mercury, and it contained 12 p.g/g mercury .

Therefore, the deformed embryo came from an egg that prob-

ably contained somewhat more than 12 p.g/g mercury.

A fourth embryo, from the 16th egg laid by a female fed 5

p.g/g mercury, had a crossed bill and an indentation in the bill

.

The estimated value of only 0 .44 p.g/g mercury in this egg has

to be viewed with caution for a number of reasons . First,

deformities were limited to the bill. Second, the estimated

mercury concentration of 0.44 was arrived at by taking the

average of a value of 0 .87 p.g/g mercury measured in egg

number 17 and a value of 0.01 for egg number 15. Egg 15 was

laid while the female was still on its uncontaminated diet ; it

was reported to contain less than the detection limit of 0 .0198

p

.g/g mercury, and consequently was assigned the value of

0.01, which is one-half the detection limit . Because of the

uncertainty of how quickly mercury would build up in the very

first egg laid after the switch to the mercury-containing diet, it

is possible that the estimate of mercury in egg 16 was some-

what on the low side, but we can say with some assurance that

this egg contained less than the 0.87 p.g/g reported for the egg

laid the following day. Finally, after the deformed embryo from

egg number 16, eggs 18, 20, 22, 26, and 28 all hatched and

eggs 24 and 30 died; therefore, this female also had an irregular

history of embryotoxic effects

.

A total of three embryos from the 20 breeding pairs of

controls also were found to be deformed. Three deformed

embryos is a low number considering that all but 4 of the 20

pairs of controls produced 60 eggs, and the lowest of these 4

G. H. Heinz and

1) . J

.

Hoffman

a The pen number of the female that laid the egg, followed by the egg number (according to the sequence in which the eggs were laid)

.

b The concentration of mercury in the egg with this deformed embryo was estimated by taking the average of the mercury in the egg laid the day

before and the egg laid the day after this egg

.

still produced 48 eggs for incubation. One of the deformed

controls had a slightly shorter lower bill than upper bill, an-

other had a shorter upper bill plus exencephaly, and the third

was a case of conjoined twins with exencephaly and a crossed

bill .

Interpreting the results for deformities is clearly not as

straightforward as for neurological signs. Though we do not

rule out the possibility that as little as I p .g/g or even less

mercury did in fact cause the deformities we saw, at the same

time we believe these estimates should be taken with some

caution.

Failure of Eggs to Hatch

The lowest estimated mercury concentration in an egg that

failed to hatch was 0.74 pg/g on a wet-weight basis and the

highest value was 38 µg/g (Table 3). In addition to the mer-

cury-treated pens listed in Table 3, 10 of the 20 control pens

also met the criterion of having no more than one of their first

12-15 fertile eggs fail to hatch . Because the odd-numbered

eggs from control pens were incubated along with the even-

numbered eggs, there were as many as 15 eggs incubated from

eggs 16 through 30. The mean hatching success of this second

set of 15 eggs from the 10 control females was 89 .2%, with a

range from 80% to 100% . In only I of these 10 control pens did

2 eggs in a row fail to hatch. The uniformly high hatching

success of eggs 16 through 30 for this select group of control

pens suggests that it is unlikely that a series of eggs from the

similarly select group of mercury-treated females in Table 3

would have failed to hatch by chance alone . Almost certainly,

most of the eggs listed in Table 3 did die from mercury

poisoning.

The lowest value in the table, 0 .74 µg/g, came from pen 84,

which hatched all 14 of its first set of eggs and all of the

ducklings survived for 6 days . Based on this pretreatment

success, we would have expected egg 16 to hatch unless

Concentration of

Mercury in Parents'

Diet (pg/g)

Female's Pen

Number-Egg

Number"

Mercury Estimated

to Have Been in Egg

(Vg/g, wet weight) b

Description of Deformity

5

16-26

3.8

Twisted toe on right foot; abnormal right wing; body generally

underdeveloped

5

62-20

3.7

Elongated, spoon-shaped upper bill ; legs and feet abnormal

;

abnormal joint on right wing

5

62-46

1 .0

Spoon-shaped upper bill; hydrocephaly; short legs ; extra toes;

abnormal wings

5

62-48

0.93

Upper bill shorter than lower; short legs; extra toes ; abnormal

right wing

10

89-24

12

Lower bill longer; eyes poorly developed ; stunted body

;

deformed skull

20

52-22

15

Left wing does not extend; foot deformed

20

65-16

1 .0

One body sharing two fused heads; exencephaly; small eyes ;

deformed bill

20

77-22

18

Exencephaly; small right eye; lower bill longer and twisted

---

Mercury Thresholds in Eggs

Table 3. Estimated mercury concentration in eggs that failed to hatch

The pen number

of

the female that laid the egg, followed by the egg number (according to the sequence in which the eggs were laid)

.

b The numerator is the number

of

eggs that hatched, and the denominator is the number of fertile eggs that were incubated

.

` The concentration of mercury in the egg that failed to hatch was estimated by taking the average of the mercury in the egg laid the day before

and the egg laid the day after this egg

.

mercury killed the embryo. The estimated mercury value for

egg 84-16 was based on the average between the value of 0.009

lcg/g for the 15th egg (mercury was below the detection limit

of 0.0185 p.g/g, and the egg was assigned a value of one-half

the detection limit) and the value of 1 .48 p.g/g reported for the

17th egg. Even if the actual concentration of mercury in the

16th egg was closer to the 17th egg than to the 15th egg, it

would still have been somewhat less than the 1 .48 p

.g/g re-

ported for egg number 17, and this would still be the lowest

value in Table 3. What is peculiar about this pen is that eggs 18,

20, and 30 hatched, and they had estimated mercury values of

8.0, 22, and 31 jig/g. The embryos in eggs 22, 24, 26, and 28

all experienced some form of embryotoxic effect . . Even-num-

bered eggs 32 through 52 all hatched, and only one duckling

(84-32) failed to survive 6 days . The return to perfect hatching

success after the cessation of mercury in the diet of female 84

suggests that the observed failures of eggs during the mercury

treatment were in fact due to the mercury the female deposited

in these eggs. However, it is still uncertain why some eggs laid

after egg 16 but still within the period when the female was

being fed mercury showed no indications of mercury poison-

ing .

If for some reason the estimated mercury level of 0.74 p

.g/g

in egg 84-16 was not the cause of death of this embryo, then the

next lowest value in Table 3 is the 3 .4 p,g/g value for egg

19-34. The fate of the eggs laid before and after egg 19-34

provide especially strong evidence that the embryo in egg

19-34 did die from mercury poisoning. Not only did all 14

eggs collected prior to the start of the mercury treatment for

this female hatch, but egg numbers 16 through 28 (eggs laid

while the female was on the mercury diet) all hatched . Egg 30,

the last laid while on mercury treatment, failed to hatch, and its

estimated concentration of mercury (based on an egg laid 3

days before and another laid 3 days after egg 30) was 4 .8 vg/g .

261

Egg 32 also failed to hatch; egg 31 was not analyzed, but egg

33 contained 4.2 p

.g/g mercury. Egg 32, being closer to the

period when mercury was in the female's diet, should have

contained slightly more than 4.2 p

.g/g. After egg 34, eggs 36

through 44 all hatched, presumably because mercury levels had

declined to a safe level in these eggs that were more removed

from the period when the female was fed mercury .

There were many other eggs from mercury-treated females

that failed to hatch and for which we have estimates of mer-

cury, but the data did not meet the rigorous criteria we im-

posed: (1) no more than one of the fertile pretreatment eggs

failed to hatch and (2) the neighboring eggs analyzed for

mercury had to have been laid no more than I day before and

I day after the egg whose mercury concentration we wished to

estimate. We did examine these data and there were no mer-

cury values lower than even the second lowest value of 3

.4

µ.g/g in Table 3 .

Some of the estimated concentrations listed in Table 3 prob-

ably were higher than what was necessary to kill those em-

bryos. For example, a series of increasing mercury levels is

seen in Table 3 in eggs 63-20, 63-26, 63-28, and 63-30 . In

addition to these embryos that died, 63-16 and 63-18 also died,

but we were only able to estimate that mercury was less than

5.0 p g/g (the level reported in egg 19). Egg 63-22 produced a

duckling that exhibited neurological signs of methylmercury

poisoning (this egg is listed in Table 1). Egg 63-24 hatched, but

the chick was unable to walk normally and died. Egg 63-30, for

example, probably did not require the 38 p.g/g mercury it

contained to kill the embryo ; it probably would have died had

its mercury level been closer to the value of 13 p g/g listed for

egg 63-20. The reason 63-30 had 38 µg/g mercury was because

it was laid later in the period when the female was being fed

mercury; this egg, therefore, probably had more than enough

mercury to kill the embryo

.

Concentration

of

Mercury in Parents'

Diet

(µg/g)

Female's Pen

Number-Egg

Number"

Fraction

of

Fertile Eggs

that Hatched During the

Pretreatment Periods

Mercury Estimated

to Have Been in Egg

(p

.g/g, wet weight)'

5

11-28

13/14

5.2

5

34-26

13/14

7.1

10

82-28

13/14

14

10

82-30

13/14

17

10

89-24

13/14

12

10

89-26

13/14

13

10

89-28

13/14

14

20

19-34

14/14

3.4

20

46-34

14/14

13

20

63-20

I1/12

13

20

63-26

11/12

29

20

63-28

11/12

30

20

63-30

11/12

38

20

65-24

12/13

23

20

84-16

14/14

0.74

20

84-26

14/14

33

20

84-28

14/14

34

---

262

Mercury in Eggs That Hatched and the Ducklings

Survived

The results in Table 4, showing high levels of mercury that

were not associated with harm, were not clear-cut. Except for

pen number 84, there was not much difference between the

mercury concentrations in eggs from the same female that were

poisoned by mercury and those not poisoned . However, results

discussed for data in the other tables suggest that less contam-

inated eggs may suffer from methylmercury poisoning,

whereas much more contaminated eggs from the same female

may hatch normally .

It addition to the differences in sensitivity that may exist

among embryos from the same parents, it is clear that embryos

from different parents can differ greatly in their sensitivity to

methylmercury. As has been mentioned, the neurological re-

sults in Table I are the most reliable of all our data because we

have never observed control mallard ducklings that exhibited

these signs. Several ducklings hatched from eggs estimated to

contain less than 5 p

.g/g mercury suffered neurological effects

.

For other ducklings, more than 20 p

.g/g mercury in the egg was

needed to cause the same effects. The same observations can be

made with the data in Tables 2 and 3 ; the concentration of

mercury associated with embryotoxic effects can differ by

more than an order of magnitude among the eggs from different

parents .

Considerations of Mercury in Egg White Versus Egg Yolk

One final consideration in interpreting our results deserves

mentioning. Because the egg-laying females in our study were

in the process of building up mercury in their bodies during the

period when mercury was fed and were losing mercury once

they were returned to an uncontaminated diet, mercury in egg

yolks and egg whites were rising and falling during these two

periods.

In our study, when the female mallards were first started on

their mercury-contaminated diets, the first eggs they laid would

have had very little mercury in the yolk because the yolk,

which takes many days to form, would have been almost

completely formed prior to exposure to methylmercury . With

these early eggs, an even greater than normal preponderance of

the mercury would be in the albumen, which forms during the

day the egg is laid. When chickens were fed methylmercury,

about 95% of the mercury in their eggs was found in the

albumen (Tejning 1967). During about the first half of incuba-

tion, an embryo relies almost completely on using the yolk for

food (Freeman and Vince 1974) . The relative contribution of

mercury in the yolk and mercury in the albumen to the toxicity

of mercury to the embryo is not known, but Tejning (1967)

found that about 10 p.g/g mercury in chicken eggs caused

embryonic death before the 10th day of incubation, which is

prior to the utilization of albumen as a food resource by the

chicken embryo (Freeman and Vince, 1974)

.

Nearly all of the deformed embryos that appear in Table 2

were well beyond the stage when they were utilizing only yolk

as a source of food, and nearly all of the embryos that failed to

hatch and are listed in Table 3 were older embryos . Further-

more, most of the embryos that suffered toxic effects were not

G. R Heinz and D . J . Hoffman

from eggs laid shortly after the switch to the mercury diets or

shortly after the switch back to the uncontaminated diet . Con-

sequently, we doubt that shifting ratios of mercury in yolk to

albumen could have complicated our interpretations very

much. However, with wild birds in nature, shifting ratios of

mercury in yolk versus albumen could be common when breed-

ing females move into and out of mercury-contaminated areas

.

Until the relative toxic contributions of mercury in the albumen

versus the yolk are understood, it cannot be known whether

basing toxic mercury thresholds on whole eggs could be mis-

leading .

Conclusions

Our current study with individual eggs provides more reliable

information than our earlier studies if one wishes to discover

the lowest concentrations of mercury in eggs that will harm the

most sensitive mallard embryos .

The findings for neurological signs, which were almost cer-

tainly related to mercury in the eggs, clearly showed that

concentrations of mercury as methylmercury in excess of 2

µg/g on a wet-weight basis will harm sensitive mallard em-

bryos. Data related to deformities and embryo mortality sug-

gest that even lower concentrations of mercury in eggs may

cause harm, but these findings were not as definitive as were

the data for neurological signs. Therefore, we believe it is

wisest to conclude that there is some evidence from deformities

and mortality, although not conclusive, that wet-weight mer-

cury concentrations of about I p

.g/g or perhaps a little below I

p.g/g can harp the most sensitive mallard embryos.

Our conclusions lend more reliability to the findings from

other mercury studies in which thresholds for effects were

expressed as mean concentrations of mercury in sample eggs

from different treatment groups. When mallards were fed 3

Itg/g mercury as methylmercury during two successive breed-

ing seasons, toxic effects on embryos and young were associ-

ated with samples of eggs that contained between about 6 and

9 p

.g/g on a wet-weight basis (Heinz 1974, 1976). These older

studies were different than the current study in that the breed-

ing adults were started on their mercury diets well before the

breeding season, and, consequently, mercury concentrations in

eggs were fairly stable over the time when eggs were collected .

Levels of mercury in eggs much below the 6-9 µg/g range

were not produced in the early studies, so the mean levels of

mercury in eggs were higher than what might have been

minimal to cause harmful effects .

In another earlier study, breeding mallards were fed 0.5 p

.g/g

mercury as methylmercury over three generations (Heinz

1979). When data for all three generations were combined, the

number of I-week-old ducklings produced per breeding pair

was significantly lower for the pairs fed 0.5 µg/g mercury than

for controls. Samples of 9-14 eggs were collected in each of

the three generations for mercury analysis . The mean levels of

mercury in eggs were 0.79, 0.86, and 0.84

sg/g

on a wet-

weight basis in the three generations, respectively . These three

means from this 1979 study were the lowest to have been

associated with reproductive impairment in mallards . Based on

the findings of that study, and lacking mercury toxicity data

specific to the wild birds they studied, some investigators doing

---

Mercury Thresholds in Eggs

Table 4. Estimated mercury concentration in eggs that hatched and the ducklings survived 6 days

263

a

The pen number of the female that laid the egg, followed by the egg number (according to the sequence in which the eggs were laid)

.

b The concentration of mercury in the egg that hatched was estimated by taking the average of the mercury in the egg laid the day before and the

egg laid the day after this egg

.

`The numerator is the number of eggs that hatched, and the denominator is the number of fertile eggs that were incubated

.

d These eggs were neighboring eggs laid by the same female that produced the eggs listed in the second column that hatched . These eggs failed

to hatch or, if noted in brackets, the embryos were deformed or exhibited neurological signs of mercury poisoning

. The estimated mercury

concentrations are shown in parentheses (a - symbol indicates that the estimate for the even-numbered egg used odd-numbered eggs that were

not always only I day removed from the egg being estimated)

.

field studies have adopted a value of about 0 .8 pg/g mercury on

bryos could be discovered . However, from a practical point of

a wet-weight basis as a best estimate of how much mercury it view, we do not think that mercury levels in mallard eggs of I

might take to harm the reproductive success of the birds they

pg/g or less are likely to harm more than a small percentage of

studied (Henny

et al.

2000; Lonzarich

et al.

1992)

.

embryos. As our data in Table 4 showed, many embryos from

The fact that findings from our recent study, designed spe-

eggs with 10 or more pg/g mercury seem to do well

.

cifically to estimate the least amount of mercury that will harm

an embryo, do not contradict the use of the 0.8 pg/g threshold,

taken from Heinz (1979), raises an interesting question. Why

did the earlier study, which was based on mean mercury levels

in groups of eggs whose individual fate was not even deter-

mined, come up with about the same threshold as the more

carefully designed current study? We believe the answer lies in

the fact that the approximate value of 0.8 pg/g from Heinz

(1979) was an average . The average probably included some

mercury concentrations that were below an effect threshold

plus some that were higher than what was needed to harp the

most sensitive embryos. Even though the older study (Heinz

1979) yielded about the same answer, the findings from the

current study, designed specifically to determine the lowest

concentrations of mercury in mallard eggs that harm the most

sensitive embryos, are more appropriate to establishing a toxic

mercury threshold in eggs that can be used as a default thresh-

old for the eggs of other species. It is important to recognize,

however, that any default threshold assumes that the embryos

of wild birds are of about the same sensitivity to mercury as are

mallards. This may not be true. When results from various lab

and wild birds were compared, there was evidence suggesting

that reproductive success of some species may be more sensi-

tive to methylmercury than it is for other species (Koster

et al.

1996) .

Our study was not designed specifically to determine what

percentages of embryos or ducklings from eggs with varying

amounts of mercury would be banned, but we did gain some

insights. The 60 pairs of mallards fed the three mercury diets

was a large number, but if one were to use even larger numbers

of breeding pairs it is possible that even more sensitive em-

Acknowledgments .

We are grateful to Carol Erwin, Laura Heinz,

Michael Hoffman, and Howard Townsend for help in caring for the

breeding pairs and gathering data

.

References

Bar JF (1986) Population dynamics of the common loon

(Gavin

immer)

associated with mercury-contaminated waters in north-

western Ontario. Can Wildl Serv Occas Pap 56, 23 pp

Eisler R (2000) Handbook of chemical risk assessment: health hazards

to humans, plants, and animals. Volume I, Metals . Lewis, Boca

Raton, FL

Fimreite N (1971) Effects of dietary methylmercury on ring-necked

pheasants. Can Wildl Serv Occas Pap 9, 39 pp

Fimreite N (1974) Mercury contamination of aquatic birds in north-

western Ontario. J Wildl Manage 38:120-131

Finley MT, Stendell RC (1978) Survival and reproductive success of

black ducks fed methylmercury. Environ Pollut 16:51-64

Freeman BM, Vince MA (1974) Development of the avian embryo

.

Chapman and Hall, London

Gile JD, Meyers SM (1986) Effect of adult mallard age on avian

reproductive tests. Arch Environ Contain Toxicol 15 :751-756

Heinz G (1974) Effects of low dietary levels of methyl mercury on

mallard reproduction . Bull Environ Contain Toxicol 11 :386-392

Heinz GH (1976) Methylmercury : second-year feeding effects on

mallard reproduction and duckling behavior . 3 Wildl Manage

40:82-90

Heinz GH (1979) Methylmercury : reproductive and behavioral effects

Concentration

of Mercury in

Parents' Diet

(pg/g)

Female's Pen

Mercury (pg/g, wet

Number-Egg

weight) Estimated to

Number for Eggs

Have Been in Eggs

that Hatched'

That Hatched"

Fraction of Fertile

Eggs That Hatched

During Pretreatment

Periodic

Estimated Mercury in Other Even-Numbered

Eggs That Suffered an Embryotoxic Effect

and Were from the Same Femaled

5

16-28

5.7

11/13

16-26 (3 .8 pg/g), 16-30

(5.7 i4g)

5

34-28

> 7.3

13/14

34-26 (7.1 pg/g)

34-30

> 7.3

10

89-30

> 14

13/14

89-24 (12 pg/g), 89-26 (13 pg/g), 89-28

(14 pg/g)

20

77-26

23

14/14

77-22 [deformed] (18 pg/g), 77-24

77-28

23

[neurological signs] (23 pg/g)

20

84-18

8.0

14/14

84-16 (0.74 pg/g), 84-22 [neurological signs]

84-20

22

(°28 4g), 84-24 [neurological signs]

84-30

31

(m29 jig/g)

---

264

on three generations of mallard ducks. J Wildl Manage 43:394-

401

Heinz GB, Fitzgerald MA (1993) Reproduction of mallards following

overwinter exposure to selenium . Environ Pollut 81 :117-122

Heinz GH, Hoffman DJ (1998) Methylmercury chloride and sel-

enomethionine interactions on health and reproduction in mal-

lards . Environ Toxicol Chem 17:139-145

Heinz GH, Locke LN (1976) Brain lesions in mallard ducklings from

parents fed methyhnercury . Avian Dis 20:9-17

Henny CJ, Grove RA, Bentley VR (2000) Effects of selenium, mer-

cury, and boron on waterbird egg hatchability at Stillwater, Mal-

heur, Seedskadee, Ouray, and Benton Lake National Wildlife

Refuges and surrounding vicinities . National Irrigation Water

Quality Program Information Report 5, Bureau of Reclamation,

79 pp

Hoffman Dl, Moore JM (1979) Teratogenic effects of external egg

applications of methyl mercury in the mallard,

Anas platyrhyn-

chos

.

Teratology 20:453-462

Koster MD, Ryckman DP, Weseloh DVC, Struger 7 (1996) Mercury

levels in Great Lakes herring gull

(Lams argentafus)

eggs, 1972-

1992- Environ Pollut 93 :261-270

Lonzarich DG, Harvey TE, Takekawa iii (1992) Trace element and

organochlorine concentrations in California clapper rail

(Rallus

longiroslris obsoletus)

eggs. Arch Environ Contam Toxicol 23 :

147-153

Meyer MW, Evers DC, Hartigan JJ, Rasmussen PS (1998) Patterns of

G. H. Heinz and D . J. Hoffman

common loon

(Gavia immer)

mercury exposure, reproduction,

and survival in Wisconsin, USA. Environ Toxicol Chem 17:184-

190

Rumbold DG, Niemczyk SL, Fink LE, Chandrasekhar T, Harkanson

B, Lame KA (2001) Mercury in eggs and feathers of great egrets

(Ardea albus)

from the Florida Everglades. Arch Environ Contain

Toxicol 41 :501-507

Scheuhammer AM, Perrault JA, Bond DE (2001) Mercury, methyl-

mercury, and selenium concentrations in eggs of common loons

(Gavia immer)

from Canada. Environ Monit Assess 72:79-94

Tejning S (1967) Biological effects of methyl mercury dicyandiamide-

treated grain in the domestic fowl

Gallus gallus L

.

Oikos (Suppl

8), 116 pp

Thompson DR (1996) Mercury in birds and terrestrial mammals . In

:

Beyer WN, Heinz GH, Redmon-Norwood AW (eds) Environmen-

tal contaminants in wildlife: interpreting tissue concentrations

.

Lewis, Boca Raton, FL, p 341

Wiemeyer SN, Lamont TG, Bunck CM, Sindelar CR, Gramlich FJ,

Frazer JD, Byrd MA (1984) Organochlorine pesticide, polychlo-

robiphenyl, and mercury residues in bald eagle eggs-1969-79-

and their relationships to shell thinning and reproduction . Arch

Environ Contam Toxicol 13 :529-549

Wolfe ME, Schwarzbach S, Sulaiman RA (1998) Effects of mercury

on wildlife: a comprehensive review. Environ Toxicol Chem

17:146-160

---

ELSEVIER

Abstract

Anthropogenic mercury (Hg) air emissions for the eight Great Lakes states in 1999-2000 were evaluated by analyzing three

inventories. The US Environmental Protection Agency (EPA) National Emissions Inventory (NEI) had the most complete coverage

for all states, and total Hg emissions ranged from 4226 lb in Minnesota to 15,928lb in Pennsylvania . Coal-fired electric utilities

accounted for 52.7% of the region's Hg emissions, varying from 20 .2% of the total in New York to 67.5% in Ohio. Other important

contributors to regional emissions included municipal waste combustion

(5.6%), mercury-cell chlor-alkali plants and hazardous-

waste incinerators (4% each), stationary internal combustion engines (ICEs)

(3.5%), industrial, commercial, and institutional (ICI)

boilers

(3.3%), and lime manufacturing (3.0%). Although medical waste incineration accounted for just over 1% of regional

emissions using the original classifications, the inclusion of health care facilities that may have been inappropriately identified with

other sectors would increase the sector to 4.5% of regional emissions (and decrease the stationary ICE sector to 1 .4% of the regional

total). There were substantial differences for some sectors between the NEI and the Great Lakes Regional Air Toxics Emissions

Inventory (GLEI), as well as unexplained differences within inventories between states (particularly for the cement, lime, and

asphalt industries, and for lamp breakage) . Toxics Release Inventory data for 2000 mainly covered electric utilities, and differences

from the NEI were significant for several states . An independent assessment indicates the possibility of underestimated Hg emissions

by about twofold for ICI boilers, although data for the sector (in particular concerning fuel oil emissions) are highly uncertain

.

Limited data indicate the likelihood of significant underestimates of electric arc furnace mercury emissions in the NEI and GLEI

inventories. Several measures are here identified for improving the reliability of the inventories, both for modeling of atmospheric

transport and deposition modeling and for tracking progress in Hg reduction initiatives

.

U 2004 Elsevier Inc. All rights reserved

.

Keylvords:

Mercury; Emissions; Inventories ; Great Lakes; Atmospheric deposition

Available online atwww.sciencedirect.com

SCIENCE

C

DIRECT-

Environmental Research 95 (2004) 282-297

Assessment of mercury emissions inventories

for the Great Lakes states

Michael Murray* and Stacie A. Holmes

l

National Wildlife Federation, Great Lakes Natural Resource Center, 213 W. Liberty St., Suite 200, Ann Arbor, M7 48104, USA

Received 20 November 2003 ; received in revised form 5 February 2004 ; accepted I I February 2004

1. Introduction

Mercury (Hg) contamination is widespread in the US

Great Lakes region. Statewide consumption advisories

due to Hg contamination covering at least one fish

species are in place in seven of the eight Great Lakes

states, covering inland lakes (Michigan, Minnesota,

Wisconsin), rivers (Indiana), or both (Illinois, Ohio,

°For the special issue devoted to papers presented at the

International Joint Commission meeting, Workshop on An Ecosystem

Approach to the Health Effects of Mercury in the Great Lakes Basin,

Windsor, Ontario, February 26-27, 2003

.

-Corresponding author. + 1-734-769-3351 .

E-mail

address:

murrayty,) nwf.org (M . Murray) .

'Present address: US Department of Agriculture, APHIS/PPQ-

EAB, 5936 Ford Court., Suite 200, Brighton, MI 48116, USA

.

0013-9351/S-see front matter © 2004 Elsevier Inc . All rights reserved

.

doi:10.1016/jxnvres.2004.02.007

Environmental

Research

hitp://www.elseviecennUlocale/envres

Pennsylvania) (US EPA, 2003a) . In addition to posing

risks to human health (National Research Council,

2000), Hg contamination in the region also threatens

certain fish-consuming wildlife, including loons and

mink (Evers et al ., 1998; Henry et al., 1998) .

Atmospheric deposition has been recognized for some

time as the major pathway by which most aquatic

ecosystems are contaminated by ongoing Hg loadings

(e .g ., Jackson, 1997; Fitzgerald et al., 1998; Schroeder

and Munthe, 1998; US EPA, 2000). For example, results

from the Lake Michigan Mass Balance Study indicated

that in the mid-1990s, 84°/n of total Hg loading to the

lake (considering atmospheric and tributary inputs) was

through atmospheric deposition (Landis and Keeler,

2002). Mercury is released to the atmosphere through

both natural processes and human activities (Gustin,

---

2003; US EPA, 1997a) . Studies of inland lake sediment

cores, however, show a significant anthropogenic Hg

signal; the modern/preindustrial Hg flux ratios in eight

rural Minnesota lakes ranged from 3 .0 to 6.7 (Engstrom

and Swain, 1997), and a similar range was found in the

Adirondack region of New York (Lorey and Driscoll,

1999). A study of Great Lakes sediment cores showed

significantly higher impact of human activity on Hg

loadings to the lakes, with increases in areal loadings of

up to 336-fold over preindustrial values in Lake

Ontario . The increased loadings were attributed to more

local industrial sources and distinct sediment mixing and

integration processes (Pirrone et al ., 1998). Although

some of the inland lake data indicate that Hg loadings in

the region peaked several decades ago, there is no

indication that loadings have returned to preindustrial

values, and anthropogenic Hg releases in the United

States and other regions of the world are still significant

(United Nations Environment Programme-Chemicals,

2002) .

There have been increasing efforts in the past decade

to develop more comprehensive Hg emissions inven-

tories. Accurate inventory data-including information

on parameters such as source location, stack height, and

Hg emissions amount and speciation-are important as

source terms for atmospheric transport and deposition

models (e.g ., Bullock and Brehme, 2002; Seigneur et al .,

2003; Cohen et al., 2004) . Accurate inventories are also

necessary in the policy context, both in assisting in

regulatory and voluntary program decision making, and

in assessing progress in meeting broader policy goals,

such as the virtual elimination targets for persistent

toxic substances in the Great Lakes Water Quality

Agreement and the Canada-US Binational Taxies

Strategy (International Joint Commission, 1987; Envir-

onment Canada and US EPA, 1997) .

Known historic Hg air emissions sources in the

United States have included coal-fired electric utilities,

certain mining practices, product- or use-related emis-

sions (e.g ., through waste disposal and certain manu-

facturing industries), and other incidental or use-related

emissions. This article presents an overview and analysis

of Hg emissions in the Great Lakes states in 1999, based

on recently released inventories for the region. The

analysis addresses most of the major Hg-emitting

sectors; more detailed analysis of coal-fired electric

utilities in the region will he presented in a separate

article (in preparation)

.

2. Materials and methods

Three inventories covering broad activity sectors and

including emissions data for Hg in the Great Lakes

states were considered in this assessment. Brief over-

M. Murray, S. A. Holmes l Environmental Research 95 (2004) 282-297

283

views of the inventories and the approach used for data

acquisition follow

.

2.1 .

National Emissions Inventory (AEI)

The NEI is an effort by the US Environmental

Protection Agency (EPA) to establish a comprehensive

inventory for both criteria and hazardous air pollutants

(HAPs). Intended uses of NEI data include in the

National, Air Toxics Assessment, in

residual risk

analysis, and in other atmospheric transport modeling

efforts (Pope et al ., 2002 ; ERG, Inc., 2003a). The first

systematic national inventory for HAPs, including Hg,

was compiled in 1990, through what was then termed

the National Toxics Inventory (NTI), and a second NTI

inventory for HAPs was completed for 1996. EPA

released the Final Version 3 .0 of the 1999 NEI in July

2003 (Pope et al., 2002; ERG, Inc., 2003a)

.

The inventory for Hg and other HAPs is divided into

point, nonpoint (NPS), and mobile sources . Point

sources include major and area sources. Major sources,

as defined in the 1990 US Clean Air Act (CAA), are

those facilities that have the potential to emit >_ 10 tons

per year of one HAP or 325 tons per year of any

combination of HAPs . Facilities whose annual emis-

sions are below these thresholds are considered area

sources. Data in the 1999 HAP inventory are derived

from a number of sources, including state, local, and

tribal agencies; industry; EPA data for regulated source

categories,

through

Maximum Achievable Control

Technology (MACT) standards; Toxics Release Inven-

tory (TRI); estimated nonpoint data for sources not

included in state, local, and tribal data; and 1996 NTI

data for sources not otherwise included . For 1999 data,

all Great Lakes states except Ohio had submitted point

source emissions data to the US EPA (Pope et al ., 2002

;

ERG, Inc., 2003a)

.

NPS emissions in the NEI consist of area sources with

smaller and/or more diffuse emissions . In part, the

definition refers to sources that the US EPA would

regulate under provisions other than sections 112 or 129

of the CAA (Pope et al., 2002). Estimates for area

sources are obtained using a 'top-down' approach, in

which national-, regional-, and state-level emissions data

are used to estimate emissions at the local (county) level

.

(Area sources where an individual emissions estimate

can be obtained, but where thee total emissions fall below

the major category thresholds just mentioned, are

considered point sources and are included in the point

source NEI.) The

hierarchy of data compilation

involved state, local, and tribal data, supplemented by

MACT NPS data, supplemented by EPA data based on

emission factors and activity data. Though passed

through quality assurance/quality control measures,

differences in estimation approaches may lead to

different estimates for the same category, limiting

---

284

between-state comparability. Among the Great Lakes

states, four (Michigan, Minnesota, New York, and

Wisconsin) submitted initial or revised NPS data to

EPA (Pope et al., 2002; ERG, Inc., 2003b). For Ohio,

emissions data for 15 HAPs, not including Hg, were

submitted from the Regional Air Pollution Control

Agency covering six counties in the southern part of the

state (ERG, Inc., 2003a). In cases where no local, tribal,

or state data are submitted, the EPA develops emissions

estimates. Finally, no Hg emissions estimates were made

for mobile sources in any jurisdiction inn the 1999 NEI

.

The data analysis approach for the NEI data was as

follows. NEI point source and NPS state data files for

1999 (in MS Access format) for the eight Great Lakes

states (Illinois, Indiana, Michigan, Minnesota, New

York, Ohio, Pennsylvania, and Wisconsin) were down-

loaded from the NEI website following finalization in

summer 2003 (see US EPA, 2003b). Queries were

developed and applied to each database file to identify

only those facilities/units (point source data) or counties

(NPS data) with reported Hg releases, based on the NEI

pollutant codes for the nine forms of Hg contained in

the 1999 NEI. Speciated Hg emissions

(i.e., elemental

gaseous Hg, gaseous divalent Hg-also known as

reactive gaseous Hg-and particulate divalent Hg) were

given only for coal- and coke-fired electric utilities in the

1999 NEI .

The approach for placing individual facilities in

sectors was based on process characterization . Source

Classification Codes (SCCs) and MACT code identifica-

tions were obtained from the NEI Lookup database

available on the NEI website (US EPA, 2003b). SCCs

were available for all records and were used initially to

categorize all facilities . For cases where the SCC was

nonspecific (e.g ., 39999999, for miscellaneous manufac-

turing industries), MACT codes, if present, were used to

assign facility categorization . Key Hg-emitting sectors

with MACT codes were stationary internal combustion

engines (ICES) (0105) ; industrial/commercial/institu-

tional

(ICI) boilers (0107); stationary combustion

turbines (0108) (which were considered with stationary

ICEs in

this analysis); integrated iron and steel

manufacturing (0305); iron foundries (0308); steel

foundries (0309); lime manufacturing (0408); Portland

cement manufacturing (0410); taconite iron ore proces-

sing (0411); asphalt roofing and processing (0418)

;

hazardous waste incineration (0801

(1-4)) ; municipal

landfills (0802); chlorine production (1403) (limited to

chlor-alkali plants in this analysis) ; medical waste

incineration

(1801) ;

municipal

waste

combustion

(1802); and electric utility boilers (1808 (1-3)) .

A very small fraction of facilities had SCCs and

MACT codes indicating two different sectors ; these were

predominantly in the stationary ICE and the ICI boiler

categories (based on MACT codes). In these cases,

facilities were placed in these two categories based on

M Murray, S.A. Holmes I Environmental Research 95 (2004) 282-297

their respective MACT codes. For facilities with SCCs

indicating ICI boiler or ICEs but a different MACT

code (again, a small number of records, typically

manufacturing sources with lower Hg emissions rather

than any of the MACT categories just listed), the

facilities were maintained in the boiler or ICE cate-

gories. Facilities having neither a specific SCC (e.g .,

39999999 for miscellaneous sources) nor a MACT code

were left in the miscellaneous category

.

2.2 .

Great Lakes Air Toxics Emissions Inventory

(GLEI)

The GLEI is a project of the eight Great Lakes states

and Ontario to compile point, area, and mobile source

emissions for toxic chemicals of concern in the region . A

steering committee made up of representatives from

each state and Ontario used a software tool and a

standardized protocol to develop emissions estimates for

each chemical . Emissions sources are organized by

Standard Industry Classification (SIC) codes, SCCs,

and Area and Mobile Source (AMS) codes . The most

recent inventory, for 1999, consists of 213 HAPs,

including Hg, and estimates foI point, area, and mobile

sources (GLC, 2003)

.

Although a standard protocol is in place, reporting

restrictions and/or unavailability of data can lead to

incomplete coverage of sources in each state . In

addition, choice of emission factors for a specific source

is made by individual state representatives based on

information on the source (including control technolo-

gies in place) (0. Cabrera-Rivera, Wisconsin Depart-

ment of Natural Resources, personal communication)

.

For this analysis, Hg emissions data (aggregated by SIC

codes and SCCs) for the Great Lakes states were

obtained directly from the authors, because sector-

specific Hg data were not yet available via the Internet

(0. Cabrera-Rivera, personal communication). Because

of the paucity of measured emission factor data for

mobile sources (GLC, 2003), Hg emissions estimates for

mobile sources were not considered in this analysis,

although estimates were made in the 1999 GLEI .

2.3. Taxies Release Inventory

The TRI was established under the US Emergency

Planning and Community Right-to-Know Act of 1986,

with the first reporting year in 1987. The requirement

originally applied to industrial facilities falling in SIC

codes 20-39; having >, 10 employees; and meeting a

threshold requirement-either manufacturing or proces-

sing a chemical in excess of 25,000lb/year, or otherwise

using a chemical in excess of 10,000 lb/year (US EPA,

2001a). Reporting requirements have since been ex-

tended to other facilities, including federal facilities in

1994 and electric utilities, coal and metal mining

---

operations (with several exceptions), and other sectors

for the 1998 reporting year. In 1999, EPA lowered

reporting thresholds for Hg and other persistent,

bioaccumulative toxic chemicals, applicable to the

2000 reporting year. The new requirement stipulated

that any facility that manufactures, processes, or

otherwise uses > 101b of Hg or Hg compounds would

be required to report (US EPA, 2001b)

.

The process involves industry submittal to EPA and

state or tribal governments of chemical data (including

amount and locations of chemicals stored on-site and

estimated release amounts, with no requirement for

measured values), compilation by EPA, and publication

of annual inventories. Data reported include releases to

air, discharges to water, land disposal, underwater

injection, and on-site and off-site transfer (US EPA,

2001a). For this study, TRI air emissions data for Hg

and Hg compounds on a state level for the Great Lakes

states were downloaded via TRI Explorer (US EPA,

2003c) and processed in MS Excel. Because the lower

reporting threshold did not take effect until the 2000

reporting year (and the large majority of Hg-emitting

facilities did not meet the previous

threshold

require-

ments for Hg), this analysis focused on TRI data from

the 2000 reporting year. (Although TRI data are

incorporated into the NEI-unless overridden by other

data-the two databases represent two different data

collection efforts .) Because of the relatively small

number of facilities reporting to TRI even with the

lower threshold, Hg emission comparisons are made

only for electric utilities and state totals

.

2.4. Additional database for coal-fired utility hg

emissions

This analysis also relied upon the 1999 Information

Collection Request (ICR) for coal-fired electric utility

steam-generating units, an effort to gather information

on Hg emissions from the units in support of the US

EPA regulatory determination for these units (US EPA,

2003d). Information derived from this effort included

utility plant configurations and pollution control de-

vices, coal Hg content, and estimated Fig emissions for

all units in the United States. This has been the only

systematic effort thus far to estimate, on the basis of

measurements, Hg emissions from coal-fired utility units

in the United States; data from this effort were

incorporated directly into the 1999 NEI. For this study,

plant and state totals for facilities in the Great Lakes

states were obtained via the EPA Air Toxics website

from the plant-by-plant mercury emission estimates file

(US EPA, 2003d)

.

Because of its breadth of coverage, the US EPA NEI

was the inventory examined in greatest detail in this

assessment. In addition, more detailed comparisons

between inventories were examined for several major

M. Murray, S. A. Holmes / Environmental Research 95 (2004) 282-297

285

sectors, and independent estimates of emissions were

considered for two sectors .

3. Results

A compilation of Hg emissions estimates for the eight

Great Lakes states from the US EPA NEI is given in

Table I . The data are presented in the format

(i .e ., point,

nonpoint) used in the NEI database . A slightly different

breakdown showing sector percentages for the Great

Lakes states is shown in Fig. I . Coal-fired electric

utilities were responsible for 52.7% of the Hg emissions

in the region in 1999, ranging in individual states from

20.2% in New York to 67.5% of the total in Ohio. This

breakdown is quite different from 2000 estimates for

Ontario, where the total fuel combustion sector was

responsible for on the order of 20% of total Hg releases

of 3028 kg (Trip et al., 2004; also see discussion in

Hagreen and Lourie, 2004) . The overall Hg emissions

picture for the Great Lakes states evolved from earlier in

the decade, at which point several other nonulility

sectors were more prominent. The remainder of this

section presents more detailed discussion of Hg emis-

sions characteristics of several major sectors

.

3.1. Electricity generation

Mercury exists naturally in coal and is released upon

combustion. Electric utilities accounted for 85.6% of

total coal consumption in the United States in 1999 (US

DOE, 1999). Variables that influence the amount of Hg

released from coal combustion include the coal rank and

concentrations of other constituents (in particular

chlorine), combustion conditions in the boiler, flue gas

temperature and composition, fly ash properties, and

pollutant controls after combustion (Kilgroe et al

.,

2002). Pollution control devices for other pollutants

(i.e .,

sulfur dioxide, nitrogen oxides, or particulate

matter, PM) can have an impact on Hg emissions from

coal-fired units, although apart from fabric filters (for

PM control),

they are not generally effective at

controlling elemental Hg (Kilgroe et al ., 2002) .

Total Hg emissions from coal-fired electric utilities in

the eight Great Lakes states are given in Table 2 for four

inventories . Although coal-fired utility boiler emissions

data in the NEI were derived directly from the EPA ICR

effort, Table 2 shows minor differences between the NEI

and ICR databases for the eight states . More substantial

differences in electric utility Hg emissions are seen in

comparing 1999 ICR or NEI data to TRI data for 2000

.

These include

-30% and

-12% lower reported

emissions in Illinois and Pennsylvania, respectively,

and -I8%

higher reported emissions in Minnesota

and Ohio, compared with NEI data. Although the

higher 2000 estimates for some states may be due to a

---

286

M. Murray, S.A. Holmes / Environmental Research 95 (2004) 282-297

Table I

Mercury emissions (lb) for Great Lakes States from US EPA NEI, 1999

Source Data drawn from US EPA, 1999 NEI database (US EPA, 2003b), based on source categories (see text for details), with point/nonpoint

source breakdown retained from EPA inventory .

NA, No data available in the database. Zero values indicate value <0 .5lb. Totals may not add exactly because of independent rounding .

"Includes only facilities with utility MACT code

.

"Includes only facilities with SCCs indicating coal-fired boilers.

`Includes stationary combustion turbines

.

° Detailed breakdown is presented in Table 4

.

`As noted in text, may include facilities that could be considered to fall within specific categories above, but for which specific SCCs or MACT

codes were lacking .

combination of increased reported emissions at indivi-

dual units and TRI coverage of additional facilities not

in the ICR database, the reasons for lower 2000

estimates for Illinois (in particular) and Pennsylvania

are not clear

.

Utility Hg emissions as reported through the GLEI

varied quite considerably from those reported through

the utility ICR, as shown in Table 2 . For New York and

Ohio, no utility Hg emissions estimates were obtained

(on an SCC basis), and for Pennsylvania, utility Hg

emissions were only 28 .1% of the ICR totals (34.1%

when based on SIC codes). Although agreement was

better for four other states, the estimated utility Hg

emissions total for Illinois was nearly double (SCC

basis) the ICR total . Although some states have more

limited ability to collect air toxics emissions data (GLC,

2003), the reasons for the considerable discrepancies in

states where data are available are not clear .

3.2 .

Industrial, commercial, and institutional boilers

ICI boilers involve controlled flame combustion and

provide thermal energy to run processes or machinery or

to produce electricity . Process heaters are units in which

combustion gases do not come into direct contact with

process gases in the combustion chamber . These boilers

and process heaters are used in a variety of ICI settings,

including oil and gas extraction and petroleum refining,

chemicals manufacturing, primary metals industries,

national security operations, health services institutions,

and educational institutions . These boilers can use coal,

oil, natural gas, or other fuels as energy sources (US

EPA, 2003e). Mercury

is

a trace constituent in the fuels

(typically as mercuric sulfide in coal and as elemental Hg

in fuel oil and natural gas) and can be released upon

combustion . On the basis of an analysis of the emissions

inventory database used in development of the MACT

Sectors

Illinois

Indiana

Michigan

Minnesota

New York

Ohio

Pennsylvania

Wisconsin

Regional total

Point sources

Electricity generation

Electric utilities

coal"

6016

4885

3094

1265

1028

7117

9961

2263

35,628

Other electricity generation

26

29

29

3

605

10

710

46

1458

ICI boilers

Coal"

149

189

133

47

86

165

294

122

1187

Other

152

57

98

33

3

25

115

66

549

Stationary ICES`

552

1344

195

39

138

0

118

0

2387

Chlor-alkali plants

-

-

1653

-

1082

2735

Cement manufacturing

1216

298

67

NA

40

38

58

9

1725

Lime manufacturing

1947

NA

NA

NA

NA

NA

42

13

•

2002

Petroleum refining

151

31

NA

NA

NA

189

318

4

692

Primary metal production

NA

NA

60

635

NA

NA

329

NA

1024

Secondary metal production

68

344

258

97

NA

0

NA

60

828

Incineration°

695

641

.

902

1413

1270

593

2289

371

8173

Other point sources`

1453

774

24

495

625

417

962

1091

5841

Nonpolnr .sources

ICI boilers

38

32

52

14

132

90

72

34

464

Residential boilers

13

26

48

68

505

60

378

64

1161

Lamp breakage

82

30

69

27

463

62

49

29

811

Dental preparation and use

61

42

53

48

74

46

44

42

410

Laboratories

80

39

65

32

118

74

79

35

522

Other nonpoint sources

8

3

10

9

13

8

10

4

64

Total point sources

12,424

8592

4861

4029

3795

10,206

15,196

5127

64,229

Total nonpomtsources

282

172

297

197

1305

340

632

208

3433

Overall total

12,705

8764

5158

4226

5100

10,547

15,828

5335

67,662

---

M.

Afurrny, S A. Holmes / Environmental Research 95

(2004)

282 297

Lamp breakage

Residential boilers

12%

7%

/

CommercialfindustrialrnafitNionaland

sewage sludge incineration

- 1 4 %

Coal-fired electric utility plants

52.7%

Medical waste incineration

1.1 %

Municipal waste combustion

5.6 %

Pdmary/secondary, iron and

.- steel manufacturing

-

1 .6 %

Petroleum refining

Other sources

8,3%

10%

Glass production

12%

Lime manufacturing

3.0

Total

mercury emissions

:

Asphalt roofing and

67,662 In

concrete -

1.7 %

Cement manufacturing

2.6

Other electricity generation

t

Chlor-alkali plants

22%

4.0%

Industrial/commerciallinstitutional boilers Stationary internal combustion engines

33%

3 .5

Fig . I . Source category breakdown of combined 1999 Hg air emissions in the Great Lakes states . Data were compiled from US EPA 1999 NET .

("Total does not add to 1110 .0% its a result of independent rounding)

.

Table 2

Coal-fired electric utility Hg emissions (lb) for Great Lakes states from four inventories, 1999

Hazardous waste incineration

4.0%

287

"Data drawn from US EPA, 1999 NEI database (US EPA, 2003h), as described in text

.

From US EPA Air Toxics website (US EPA, 2003d), plant-by-plant mercury emission estimates' http .//www.epa-gov/ttn/atw/combust/utiltox/

stxstote2 .pdf file .

`Year 2(100 data, for Standard Industrial Classification (SIC) codes 4911, 4931, 4939, from US EPA TRI Explorer (US EPA, 2003c)

.

a 1999 Great Lakes Air Toxics Emissions Inventory, categorization based on SCC (data derived from spreadsheet provided by O . Cabrera-Rivera,

personal communication)

.

eSame us note d, for categorization based on SIC codes 4911, 4931 . Note that SIC codes 4911, 4931, 4939 are not exclusively for coal-fired

electricity generation, but it is assumed that most Hg emissions are associated with coal-fired units .

standard for ICI boilers by the US EPA, -42% of the

inventoried ICI boilers and process heaters in the United

States in the late 1990s were in the Great Lakes states

(US EPA, 2003t)

.

Table 3 indicates Hg emissions estimates from ICI

boilers in the Great Lakes states based on the NEI and

GLEI . Data for the NEI show that for every state but

New York, coal-fired ICI boilers were the largest

subcategory of Hg emissions in the NEI database . The

data also show that emissions from oil-tired boilers and

boilers burning gas or other fuels were also relatively

substantial in most states (as categorized in this

assessment, the "other" category may include boilers

burning a mix of coal and other fuels) . Data from the

GLEI show agreement within a factor of 2 for !CI Hg

totals with data from the NEI database for four of the

five Great Lakes states with available data. However,

coal-fired emissions for Illinois boilers from the GLEI

were eight times higher than the estimate derived from

the NEI database

.

The third set of data in Table 3 indicate Hg emissions

estimates for ICI

boilers based on statewide fuel

Inventory

Illinois

Indiana

Michigan

Minnesota

National Emissions Inventory"

6016

4885

3094

1265

Electric utility information collection request °

5989

4884

3083

1265

Toxics Release Inventory (2000)`

4169

5736

3010

1497

GLEI (SCC)d

11,868

4383

2612

1452

GLEI (SIQe

12,006

4395

2662

1531

2000 TRI % difference from ICR

-30.4

17.4

-2 .3

18 .4

New York

Ohio

Pennsylvania

Wisconsin

1028

7117

9961

2263

1027

7109

9959

2264

1175

8392

8780

2114

NA

NA

2800

2284

100

NA

3400

2577

14.4

18.0

-11,8

-6.6

---

288

M. Murray, S.A .

Holmes I

Environmental

Research 95 (2004) 282-297

Table 3

ICI boiler Hg emissions (Ib) for Great Lakes states from two inventories and fuel consumption approach, 1999

consumption data and emissions factors (derived for the

US EPA MACT standard development process) for coal

and fuel oil (see footnotes c and d in Table 3). In

deriving these estimates, it was assumed that all such

fuel in each category was consumed at ICI boilers in

each state .

For coal

boilers, data from

the fuel

consumption approach are uniformly higher than NEI

estimates for each state (ranging from 1 .6- to 4.3-fold

higher than the NEI estimates, for Pennsylvania and

Minnesota, respectively), and higher for two of the five

states with available data in the GLEI ; the total regional

emissions in the fuel consumption approach are 2 .3-fold

higher than the total estimate for all eight states in the

more complete NEI. For oil-fired boilers, there are even

greater discrepancies, with the fuel consumption esti-

mate/inventory estimate ratio ranging from 0.3 (Illinois

in the NEI) to 14.7 (Minnesota in the GLEI), and the

fuel consumption approach overall leading to a nearly 2-

fold higher estimate than the NEI for the eight states

combined. Relative to other states, it is likely that the

NEI estimate for New York oil-fired boiler emissions in

particular is low; whereas the state's oil-fired ICI boiler

Hg emissions in the NEI amounted to 22%

of regional

emissions for the sector in 1999, the state accounted for

Overall totals may not add exactly due to independent rounding .

"Data drawn from US EPA NEI database (US EPA, 2003b), as described in text. Space heaters and boilers with SCCs not specifically indicating

coal or

oil

(i.e .,

102001xy-102005xy for industrial, and 103001 xy and 103002xy for commercial/institutional) are included in gas and other category .

1999 Great Lakes Air Taxies Emissions Inventory data. Same category breakdown as for NEI (data derived from spreadsheet provided by (y

Cabrera-Rivera, personal communication)

.

`Statewide coal consumption data for nonutility industrial plants are from DOE (1999) (Table 71, with estimate for New York based on 1998 and

1999 data for three states in that census region); emission factor of 5.43 lb per trillion BTUs from ERG, Inc . (2002) (Appendix A); heat content of

coal from DOE (1999) (Table 107). Mercury emissions calculated as product of coal consumption data (converted to BTU basis from heat content

data) and emission factors .

d Statewide oil sales data for industrial and commercial sectors are from DOE (2000), with adjusted distillate and residual fuel

oil

data from Tables

16, 17, respectively; average heating values for distillate and residual oils of 136,725 and 152,400 BTUs/gal, respectively, derived from US EPA

(1997b) (Table 6-10, averaging No

.

1,

No. 2 fuel

oil

values (distillate) and No .6 (residual) for central and eastern regions for each fuel); Hg emissions

factors of 0.00887 and 8.801b per trillion BTUs, respectively, from ERG, Inc. (2002, Appendix A). Mercury emissions for distillate and residual fuels

calculated as product of oil consumption data (converted to BTU basis from heat content data) and emission factors

. Totals for both fuels are shown

in table.

>65% of industrial and commercial sector residual oil

sales in the eight states (US DOE, 2000) .

Discrepancies in Hg emissions estimates for the three

approaches could be due to a number of variables

(including different coverage in number

of

boilers

between each database), but one important issue is

choice of emission factors. For coal combustion, the

latest version of the US

EPA Factor Information

Retrieval Data System (FIRE 6.23) has Hg emission

factors for bituminous/subbituminous-fired ICI

boilers

of 16 and 3 .21b per trillion BTUs of heat input, for

uncontrolled

and controlled

emissions,

respectively

(US EPA, 2003g; the latter converted based on average

heating values in US EPA, 1997b). The average value

derived by the US EPA in the

ICI boiler MACT

database (used as well in the fuel consumption approach

presented

in Table

3),

which

takes

into account

that some facilities

have pollution control

equip-

ment installed, is 5 .43 lb per trillion BTUs (ERG, Inc .,

2002). This value is in the range of median values

derived from the utility

ICR for bituminous and

subbituminous coals (7.1 and 5 .Olb Hg per trillion

BTUs, respectively) (Kilgroe et al ., 2002) . If emission

factors similar to the lower value from the

FIRE

Inventory/sector

Illinois

Indiana

Michigan

Minnesota

New York

Ohio

Pennsylvania

Wisconsin

Regional total

NEI"

Coal

160

194

134

48

98

169

301

129

1232

Oil

58

35

65

14

116

87

132

27

533

Gas and other

121

49

84

33

7

25

48

67

434

Total

339

278

283

94

222

280

481

223

2199

GLEI"

Coal

1279

89

133

43

NA

NA

NA

220

1764

Oil

34

9

14

2

NA

NA

NA

8

66

Gas and other

83

42

57

33

NA

NA

NA

151

368

Total

1396

140

205

78

NA

NA

NA

379

2198

Estimate based on statewide fuel consumption

205

203

453

496

212

2822

Coal'

449

540

264

Oild

17

23

17

31

645

80

109

66

987

Total

466

563

281

236

849

533

604

278

3809

---

database

(3.21b

per

trillion BTUs)

were

used

in

development of the NEI and GLEI on a considerable

number of coal-fired ICI boilers in the region, this could

partly explain the lower estimated Fig emissions in the

inventories compared with the fuel consumption ap-

proach

.

For petroleum products, the uncertainties surround-

ing emission factors are even higher than for coal . A US

EPA compilation of estimated Hg concentrations in

crude oils from various regions of the world gave mean

values ranging from 0.1 to 1505 parts per billion (ppb),

and mean Hg concentrations for gas condensates

ranging from 15 to 3964pph (US EPA, 2001c). A

review of petroleum-related Hg emissions indicated

mean Hg levels in distillate and residual fuel oils ranging

from 4 to 400 ppb (Wilhelm, 2001)

.

These wide-ranging values have been mirrored in US

EPA emission factors derived over the past decade

.

Whereas a value of 7 .0 lb Fig per trillion BTUs for

residual and distillate fuel oil was used to estimate

commercial and residential boiler Hg emissions in the

US EPA Mercury Study Report to Congress, an

emission factor of 0.616 Hg per trillion BTUs was

derived from measurements on utility residual fuel oil

(US EPA, 2001c). The latest version of the US EPA

FIRE emission factor database has a distillate oil Hg

emission factor (uncontrolled) of 3 .01b per trillion

BTUs of heat input, or intermediate between the two

values above, while no value was available for residual

fuel oils (US EPA, 2003g). As part of its ICI boiler

MACT standard development, the US EPA derived

emission factors for residual and distillate-fired boilers

of 8.8 and 0.00887 lb Hg per trillion BTUs, respectively

(ERG, Inc., 2002). The distillate factor is nearly 340

times lower than the distillate oil Hg emission factor in

the FIRE database, whereas the residual factor is at the

high end of published emission factors . These latter two

factors were used to obtain the emissions estimates in

the fuel consumption approach given in the last section

of Table 3, and (given the emission factor differences)

most of the Hg emissions were associated with residual

oil . Assuming that the emissions factors from the ICI

boiler MACT standard development are in fact more

appropriate for the Great Lakes boiler population, the

net effect (of a lower distillate factor but a high residual

oil factor) may indicate underestimated total oil-fired

ICI boiler Hg emissions in

the NEI and GLEI

inventories .

However, there is a crucial need to obtain more

accurate data on Hg emissions from a wider variety of

petroleum sources. A recent article reporting analyses of

standard reference materials using a closed-system

combustion

technique

with

cold-vapor

inductively

coupled plasma mass spectrometry analysis noted the

importance of low blanks in analyzing fig concentra-

tions at the low end of the scale in petroleum products

.

M. Murrny, S.A. Holmes / Environmental Research 95 (2004) 282-297

289

These authors found a substantially higher Hg level in

residual fuel (3460pg/g) than distillate fuel, but this

value was still at the low end of other published

data (Kelly et al., 2003) .

Converting to a

heat

content basis (following the approach of US EPA

(2001c), such an Hg concentration would correspond to

an emission factor of -0.21b Hg per trillion BTUs,

or over an order of magnitude lower than most emission

factors most recently used by the US EPA . If this

finding is more typical for residual oils, and if a typical

distillate oil emission factor is indeed much lower than

for residual oils-as found in ERG Inc. (2002)-Hg

emissions associated with ICI oil-fired boilers in the

Great Lakes region would be substantially less than

indicated in the last row of Table 3 (assuming complete

coverage of all such boilers). Such an assessment

would obviously have to be confirmed with Hg

concentration data on residual and distillate fuel oil

known to he burned in a representative fraction of these

boilers

.

Finally, although the statewide fuel consumption

approach yielded higher estimated total Hg emissions

from ICI boilers than the two inventories examined,

the regional estimate based on fuel consumption is

still lower than what would be expected based on

findings from the MACT regulatory development

process (indicating -26,00016 for all fuels for the

entire United States, with no Hg emissions from

natural gas--fired boilers reported) (ERG, Inc ., 2002)

.

This is particularly puzzling given that the MACT

regulatory development process reported a considerable

number of ICI boilers in the Great Lakes region

;

for example, analysis of the inventory indicated that,

on a boiler-per-segment basis (taking into account

different fuels burned, but not boiler capacity), >55%

of the coal-fired boilers were in the Great Lakes states

(US EPA, 20030 . (Though not published in the

literature, an assessment of regional boiler emissions in

this database has been reported; Delta Institute, 2002 .)

The reasons for the lower than expected regional Hg

emissions (or conversely the higher than expected

national emissions from ICI boilers, considering (hat

the fuel consumption approach yielded regional esti-

mates within a factor of 2 of the NEI inventory estimate)

are not clear. Issues to investigate would include

emission

factors,

boiler population, capacity,

and

operational data

.

An additional oil-related source of Hg emissions is

petroleum refining. As shown in Table I, estimated

emissions varied considerably among the states, with no

data available for Michigan, Minnesota, and New York,

and highest emissions in Pennsylvania (318lb) . Much

better information on the Hg content of crude oil, as

well as ensuring complete coverage of all refining

facilities, is necessary to better quantify Hg emissions

from the sector

.

---

290

M. Murray, S. A. Holmes / Environmental Research 95 (2004) 282-297

3.3 .

Stationary ICEs and combustion turbines

Stationary ICEs range in size from 1 horsepower to

> 10,000 horsepower, and serve to generate mechanical

or electrical power at fixed sites (US

EPA, 2003h) .

Mercury can be present at trace levels in the fuels (e.g .,

gasoline, diesel oil, natural gas)

and released upon

combustion . Combustion turbines work similarly,

by

burning fuels (typically natural gas), and produce

electricity, heat, or shaft power. Combustion turbines

are used by the electric power industry,

independent

power producers, the gas pipeline industry, and chemi-

cal and industrial plants, with sizes ranging from

I to

-200MW (US EPA, 2003i) .

For purposes of this

analysis, the two sectors were considered together

.

Table 1 indicates that Hg emissions from stationary

ICEs and combustion turbines ranged

substantially,

from <0.05lb in Ohio to 13441b in Indiana, and the

total for the eight states accounted for 3 .5%

of the

region's Hg emissions. There was an equally wide range

in the number of emissions records, from 8 in Ohio to

905 in Illinois . However, inspection of SIC, emission

unit, and process codes indicated that 61 .6% of the

emissions for the sector in all states were attributable to

health care facilities, mostly from emissions at facilities

in Indiana (13331b) and Pennsylvania (1181b)

. Con-

sidering the importance of these emissions, two possible

explanations for these results are (1) mistaken placement

of the facilities in the stationary ICE category rather

than the MACT and SCC medical waste incineration

category, or (2) if categorized appropriately as ICEs,

erroneously high Hg emissions estimates . This is an issue

that is easily resolved, and because of the magnitude of

Table 4

Mercury incinerator emissions (lb) for Great Lakes states from two inventories

the emissions estimates and the likelihood

of signifi-

cantly differing Hg control requirements for ICEs

as

compared with existing requirements for medical waste

incinerators (MWIs), it is obviously important for these

facilities to be properly categorized

.

If these facilities

were indeed misclassified,

the stationary ICE Hg

emissions would drop to 9171b in

1999, or 1.4% of

the regional total

.

3.4. Waste incineration

In the mid-1990s, waste incineration accounted

for

34% of the 158 tons of anthropogenic Hg emissions in

the United States, according to the US EPA (1997a)

.

Estimation of incinerator emissions in 1999 is particu-

larly challenging, because it was a period of finalization

and/or state implementation of federal emissions stan-

dards for municipal waste combustors (MWCs), MWIs,

and hazardous waste incinerators (HWIs)

(e.g ., US

EPA, 2000). Given this caveat, estimates for the Great

Lakes states in 1999, derived from the NEI inventory,

indicate that Hg emissions

from incineration were

almost certainly down from the mid-1990s,

but were

still appreciable, at 81741b, or 12 .1% of the regional

emissions total (see Table 4 and Fig

.

1). MWCs were

most prominent, at 5.6% of the regional total, followed

by HWIs (4.0% of the total)

.

Table 4 also shows

the significant variation

in

incinerator Hg emissions among

states for a given

sector. For example, Hg emissions from MWCs in the

NEI ranged from none in Ohio to just over 1000 lb in

Minnesota, New York, and Pennsylvania

.

Emissions

from commercial and industrial solid-waste incinerators

NA, Data not available in databases . Zero values indicated reported values <0-51b. Totals may not add exactly because of independent rounding

.

Data drawn from US EPA NEI database

(US

EPA, 2003b), as described in text

.

1999

Great Lakes Air

Toxics

Emissions Inventory, categorization based on SCC (data derived from spreadsheet provided by 0. Cabrera-Rivera,

personal communication) .

Sectors

Illinois

Indiana

Michigan

Minnesota

New York

Ohio

Pennsylvania

Wisconsin

Total

NEI"

Municipal waste combustion

.28

24

536

1001

1047

NA

1002

132

3770

Medical waste incineration

51

244

50

2

38

17

283

36

721

Sewage sludge incineration

NA

NA

NA

379

3

81

14

NA

476

Hazardous waste incineration

128

372

316

32

182

494

990

194

2709

Commercial and industrial inclusive .