EPA/600/R-11/066F
October 2013
www.epa.gov
Fish Tissue Analysis for Mercury and PCBs from a New
York City Commercial Fish/Seafood Market
Prepared by:
U.S. Environmental Protection Agency
A collaborative effort between Region 2 and
ORD National Center for Environmental Assessment
ORD National Risk Management Research Laboratory
ORD National Research Exposure Laboratory
This is a Regional Applied Research Effort (RARE)
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
Preferred Citation:
U.S. Environmental Protection Agency (EPA). (2013) Fish Tissue Analysis for Mercury and PCBs from a New
York City Commercial Fish/Seafood Market, Washington, DC; EPA/600/R-11/066F. Available from: National
Technical Information Service, Springfield, VA, and online at: http://www.epa.gov/.
11
-------
TABLE OF CONTENTS
LIST OF TABLES v
LIST OF FIGURES vi
ABBREVIATIONS vii
AUTHORS, CONTRIBUTORS, AND REVIEWERS viii
ACKNOWLEDGEMENTS xi
EXECUTIVE SUMMARY xiii
1. THE COMMERCIAL FISH MARKET STUDY 1
1.1. SAMPLING METHODS 1
1.2. ORGANIZATION OF THIS REPORT 4
2. RESULTS: MERCURY CONCENTRATIONS ACROSS SPECIES 5
2.1. MARKET, COMMON, AND SCIENTIFIC NAMES 5
2.2. MERCURY CONCENTRATIONS BY MARKET NAME 9
3. MERCURY CONCENTRATION BY SPECIES, LOCATION, CONDITION, AND
SIZE 16
3.1. SPECIES 16
3.2. WATER BODY OF ORIGIN 20
3.3. EFFECTS OF BODY WEIGHT AND LENGTH ON MERCURY
CONCENTRATIONS 23
3.4. WILD VS FARMED RESULTS 23
4. ANALYSIS OF MEASUREMENT VARIABILITY 25
5. COMPARISON OF CM DAT A TO RISK METRICS 34
5.1. FISH ADVISORIES AS A COMMUNICATOR OF RISK 34
5.2. COMPARISON OF CM CONCENTRATIONS TO ACTION LEVELS 36
5.3. ESTIMATES OF THE NUMBER OF SERVINGS PER WEEK OF CM
FISH FOR ADULT FEMALES OF CHILD-BEARING AGE 39
6. COMPARISON OF CM DATA TO FDA MONITORING DATA 45
7. QUALITY ASSURANCE/QUALITY CONTROL PROCEDURES 52
7. l.QA/QC OF THE DATA DELIVERY 52
7.2. QA/QC OF THE ANALYSIS 54
8. REFERENCES 56
APPENDIX A. DETAILED STATISTICAL ANALYSIS RESULTS A-l
APPENDIX B. SUMMARY OF DNA BARCODING ANALYSIS B-l
APPENDIX C. PCB ANALYSIS C-l
in
-------
TABLE OF CONTENTS (continued)
APPENDIX D. CALCULATIONS FOR NUMBER OF SERVINGS D-l
APPENDIX E. EXTERNAL PEER REVIEW OF EPA'S DRAFT REPORT, FISH
TISSUE ANALYSIS FOR MERCURY AND PCBS FROM A NEW YORK
CITY COMMERCIAL FISH/SEAFOOD MARKET AND EPA'S RESPONSE
TO COMMENTS E-l
IV
-------
LIST OF TABLES
1. Crosswalk of Market Names to Scientific Names 6
2. Statistical Information by Market Name, Non-detects Equal to Half the Reporting
Limit 12
3. Examination of the Effect of Outlier Data in the Tuna Species 15
4. Mercury Concentrations by Species or by Condition (e.g., Hard- or Softshell)
Within a Market Name; Non-detects Equal to Half the Reporting Limit 17
5. Mercury Concentrations by Water Body of Origin Within a Market Name
Species; Non-detects Equal to Half the Reporting Limit 21
6. Variance Within Individual Composite Sample Duplicates 29
7. Variance Within Individual Composite Sample Replicates 31
8. Variability Analysis of Composite Sample Duplicates 32
9. Variability Analysis of Composite Sample Replicates 33
10. Variability Analysis of Tuna Replicate and Duplicate Groups 33
11. International, U.S. Federal, and U.S. State MeHg/Hg Action Levels 37
12. Estimated Number of Fish Servings per Week for an Adult Female of Child-
bearing Age Based on Means and Upper 95% Confidence Limits on Mercury
Concentrations by Species 42
13. Estimated Number of Fish Servings per 30 DAY MONTH for an Adult Female of
Child-bearing Age Based on Means and Upper 95% Confidence Limits on
Mercury Concentrations for Three High Mercury Species 44
14. Crosswalk Between CM Market Name and FDA Monitoring Name 46
15. Comparisons of Commercial Market Measured Concentrations to FDA
Monitoring Data 48
-------
LIST OF FIGURES
1. Mean mercury concentrations ± two estimated population standard deviation for
all fish market names and typical action levels 38
2. Bar charts comparing the commercial market mercury concentrations to the FDA
monitoring data concentrations: Bars are sample means, error limits are one
standard deviation above the mean 50
VI
-------
ABBREVIATIONS
AMF Asian market fish
BLAST Basic Local Alignment Search Tool
BOLD Barcode of Life Database Systems
CBOL Consortium for the Barcoding of Life
CM commercial market
COXI cytochrome c oxidase subunit I
CV coefficient of variation
EPA U.S. Environmental Protection Agency
EU European Union
FDA U.S. Food and Drug Administration
FFM Fulton Fish Market
HANES Health and Nutritional Examination Survey
MeHg methyl mercury
NYCDOHMHNew York City Department of Health and Mental Hygiene
PCB polychlorinatedbiphenyls
PCR polymerase chain reaction
QA/QC Quality Assurance/Quality Control
QAPP Quality Assurance Project Plan
r\
R correlation coefficient
RfD reference dose
SD standard deviation
SOP standard operating procedures
TEF toxic equivalency factors
TSS total sum of squares
vn
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
PRINCIPAL INVESTIGATOR
Mark Maddaloni, DrPH
Region 2 - New York City
U.S. Environmental Protection Agency
PROJECT MANAGER
Cheryl Itkin, MA
ORD National Center for Environmental Assessment - Washington, DC
U.S. Environmental Protection Agency
AUTHORS
Mark Maddaloni, DrPH
Region 2 - New York City
U.S. Environmental Protection Agency
Dennis Santella
Region 2 - New York City
U.S. Environmental Protection Agency
Cheryl Itkin, MA
ORD National Center for Environmental Assessment - Washington, DC
U.S. Environmental Protection Agency
Henry Kahn, DSc
ORD National Center for Environmental Assessment - Washington, DC
U.S. Environmental Protection Agency
Stan Stephansen
Region 2 - New York City
U.S. Environmental Protection Agency
Moses Chang
Region 2 - New York City
U.S. Environmental Protection Agency
Michael Borst
ORD National Risk Management Research Laboratory - Edison, NJ
U.S. Environmental Protection Agency
Vlll
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
AUTHORS (continued)
John Bourbon, MS
Region 2 Laboratory - Edison, NJ
U.S. Environmental Protection Agency
Fred El sen
Region 2 Laboratory - Edison, NJ
U.S. Environmental Protection Agency
John Martinson, MS
ORE) National Exposure Research Laboratory - Cincinnati, OH
U.S. Environmental Protection Agency
Maureen O'Neil, MURP
Region 2 - New York City
U.S. Environmental Protection Agency
Cara Henning
ICF International - Durham, NC
REVIEWERS
U.S. EPA Internal Reviewers
Gina Ferreira, MS
Region 2 - New York City
U.S. Environmental Protection Agency
Jacqueline Moya, BS
ORD National Center for Environmental Assessment - Washington, DC
U.S. Environmental Protection Agency
Maria D. Smith, MS
Office of Water - Washington, DC
U.S. Environmental Protection Agency
Rita Schoeny, PhD
Office of Science Policy - Washington, DC
U.S. Environmental Protection Agency
IX
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
REVIEWERS (continued)
External Reviewers
Henry A. Anderson, MD
Private Consultant
Madison, WI
Gary A. Pascoe, PhD, DABT
Pascoe Environmental Consulting
Port Townsend, WA
Alan H. Stern, DrPH, DABT
Independent Consultant
Metuchen, NJ
-------
ACKNOWLEDGEMENTS
This report was prepared for the U.S. Environmental Protection Agency by ICF
International under contract. Additionally, we gratefully acknowledge the programs and
individuals mentioned below for their contributions to this project.
The Commercial Fish Market Study was funded through the Regional Applied Research
Effort (RARE) Program administered by EPA's Office of Research and Development (ORD).
RARE is one approach EPA takes to promote collaboration between the Regional Offices and
ORD. The goals of the program are to:
• Provide the regions with near-term research on high-priority, region-specific
science needs;
• Improve collaboration between regions and ORD laboratories and centers; and
• Build a foundation for future scientific interaction.
These goals were met through close collaboration between various ORD and Region 2
staff. ORD and Region 2 laboratories located in Edison, New Jersey worked cooperatively to
perform sampling and tissue analysis, respectively. ORD's National Exposure Research Lab
(NERL) in Cincinnati performed DNA sequencing analysis on the samples in order to assess
concordance between market names and taxonomic classification. Carrie Drake, Dynamac
Corporation and Stephen Morris, Student Services Contractor contributed to this effort. Thomas
O'Connor (EPA ORD NRMRL), Carolyn Esposito (EPA, ORD NRMRL) and Carol Lynnes
(EPA, Region 2) provided valuable input to the Quality Assurance Project Plan (QAPP) for the
Commercial Fish Market Study. ORD's National Center for Environmental Assessment
(NCEA) provided statistical consultation, project management, and funding and contract
oversight for data analysis and the development of this report. Overall project quality assurance
was managed by Beverly Comfort (ORD NCEA).
Input on study design was provided by Michael Bolger of the Food and Drug
Administration (FDA). Numerous individuals, some now retired, from the New York State
Department of Agriculture and Markets provided valuable sampling and analytical assistance.
Those individuals included Joe Colby, Curt Vincent, and Steve Stitch. Without the benefit of the
New York City Department of Health and Mental Hygiene's (NYCDOHMH) Health and
Nutritional Examination Survey (HANES), there would have been little impetus to conduct this
study. There are many at the NYCDOHMH who assisted in this study, but none more so then
Wendy McKelvey. Wendy was involved in both the New York City HANES and a
NYCDOHMH companion study of mercury and PCBs in seafood obtained from markets in NYC
Asian communities. Wendy provided an enormous amount of input on all aspects of the study
design and execution and we would like to think that we repaid her in some small way by
XI
-------
making the services of Region 2's resident ichthyologist, Moses Chang, available to
NYCDOHMH to assist in their companion study.
Finally, this study enjoyed a considerable logistic benefit in the form of the New Fulton
Fish Market which allowed for "one stop shopping" even though it required three separate trips
to acquire the large sample size. The Market Manager, George Maroulis, Director of Security,
Ken Klein, and Coop President, Warren Kremin, provided the study team with valuable insights
and advice for negotiating our way around a very busy commercial fish market in the wee hours
of the morning. In addition, individual fishmongers, too numerous to mention by name, were
very generous with sharing their expertise on various aspects of commercial fish markets and the
products they purvey. Also we would like to acknowledge PARS Environmental, Inc. for
sample acquisition.
xn
-------
EXECUTIVE SUMMARY
The New York City Commercial Market (CM) Seafood Study was undertaken by the
Environmental Protection Agency (EPA; the New York Regional Office in collaboration with
the Office of Research and Development) to measure mercury (Hg) concentration in the seafood
most commonly consumed by residents of the New York City metro area. The goal of this study
was to obtain objective information and descriptive statistics on the levels of mercury found in
commonly consumed seafood species. The data collected was presented in a manner that allows
for informed choices of seafood consumed.
The CM study is one of two complimentary studies conducted as part of an effort to
understand and respond to the results of a Health and Nutritional Examination Survey (HANES)
conducted by the New York City Department of Health and Mental Hygiene (NYCDOHMH).
The NYC HANES included measurements of blood Hg concentration in a probability sample of
1,811 New Yorkers selected to represent the age, gender, and ethnic composition of the adult
population (McKelvey et al., 2007). The geometric mean blood Hg (approximately equal to the
median) concentration was elevated threefold compared to national estimates. Asians registered
unusually high blood Hg, with Chinese New Yorkers registering a geometric mean almost three
times that of the overall sample value. Seventy-two percent of Chinese New Yorkers in the
NYC HANES had blood Hg attaining the New York State reportable level of 5 jig/L or above,
although this is based on a small sub-sample. Citywide, NYC HANES estimated that 1.4 million
NYC adults have blood Hg at or above the reportable level. The NYC HANES survey recorded
data on the number of meals that included fish but not the amount or type offish consumed.
An Asian Market Fish (AMF) study was conducted by the NYCDOHMH. The AMF
study measured levels of Hg and polychlorinated biphenyls (PCBs) in 282 specimens of 19
species commonly sold in markets that serve the Asian community. Mean Hg levels ranged from
below the limit of detection (0.004 mg/g) in tilapia to 0.229 mg/g in tilefish. The highest Hg
level (1.150 mg/kg) was measured in a tilefish specimen. Tilefish, canned eel, blackfish, and
Spanish mackerel had the highest mean Hg levels. Porgy, yellow croaker, and Buffalo carp were
identified as fish with the highest mean PCB levels. The AMF study used the U.S. EPA
Reference Dose (RfD) for methyl mercury (MeHg) to calculate the number of 6 oz (170 g) meal
servings that a 60 kg women could eat per week (McKelvey et al., 2010).
In the CM study conducted by EPA, samples of 33 commonly consumed species were
obtained from the New Fulton Fish Market (Bronx, NY), the largest commercial seafood market
in the nation and the source of most of the fresh seafood consumed in the NYC area. Samples
from the targeted species list were purchased from vendors operating in the market. For each
species selected, multiple specimens (typically three) from the same vendor were combined to
form composite samples. All samples were analyzed for total Hg concentrations and PCB
measurements were obtained in a limited subset of the samples. Mean Hg concentrations across
composite samples for the same market name ranged from as low as 0.0054 mg/kg (shrimp) to as
high as 0.42 mg/kg (tuna). The species with the four highest Hg concentrations were tuna,
Xlll
-------
swordfish, Spanish mackerel, and mahi-mahi, while shellfish tended to have the lowest
concentrations.
Measured Hg concentrations were compared to three actions levels (Maine;
Florida/European Union [EU]/Canada; U.S. Food and Drug Administration [FDA]) and EPA's
level of concern. The derivations and uses of these levels are discussed in this report. None of
the measured Hg concentrations in the individual composite samples for any species had
concentrations that were higher than the FDA action level; however, 70% or more of tuna and
swordfish composite samples exceeded the Maine Action level, the EPA Level of concern, and
the Florida/EU/Canada Action Level, and the overall tuna and swordfish mean values exceeded
the Maine Action Level and the EPA level of concern.
As was done by NYCDOFDV1H, EPA used the concentration measurements obtained in
this study to estimate mercury intake by adult women of child-bearing age resulting from
assumed amounts offish consumption. Women of child-bearing age were used as the basis of
the calculation of Hg intake because they are, as a group, more sensitive than the overall adult
population. Thus, values considered protective for this group would be protective of the overall
adult population. The permissible daily intake of MeHg was calculated and compared to MeHg
intake from fish ingestion. The estimated amount of Hg intake was converted to a number of
servings per week that would yield a safe daily intake level of Hg. The number of servings per
week for adult women were estimated assuming a body weight of 65 to 67 kg and a serving size
of 8 oz fresh weight (or 6 oz cooked weight), and using the U.S. EPA RfD for MeHg of 0.1
Hg/kg-day. In addition, it was assumed that a woman eats only a single seafood species in a
week.
The CM market species mean composite concentrations were compared to mean
concentrations in FDA monitoring data collected from 1995-2004. This comparison showed
that the CM species mean concentrations tended to be lower than FDA concentrations.
However, the CM mean 0.42 mg/kg for tuna Hg concentrations is within 5% of the more recent
(2000-2004) FDA monitoring value (i.e., 0.40 mg/kg). A limited subsample (N = 50) across
five species (salmon, crab, tuna, catfish, and mackerel) was also analyzed for 124 different PCB
congeners. The PCB analysis was constrained by cost and detailed statistical analysis was not
performed on the resulting data which are limited. The species selected for PCB analysis was
purposeful in the sense that those species that were included are those suspected of having
elevated PCB levels. NYCDOHMH HANES did not analyze biomonitoring data for PCBs.
Overall, the composite samples selected for PCB analysis had concentrations within the FDA
tolerance level.
This study also made use of recent advances in DNA sequencing technology. "DNA
barcoding" has emerged as a useful taxonomic tool that can help overcome some of the issues
associated with morphology based identifications. Barcoding uses a short genetic sequence from
a standard part of the genome in an attempt to accurately assign a specimen to a given taxon, or
ideally, a species. Such an assignment can be made by examining a genomic region that exhibits
a high degree of sequence conservation within a species, but appreciable divergence compared to
xiv
-------
other species. DNA sequencing of a portion of the cytochrome c oxidase subunit I (coxl) gene
was performed by EPA's Office of Research and Development laboratory in Cincinnati.
Overall, there was concordance between the DNA-based results and the market names.
xv
-------
1. THE COMMERCIAL FISH MARKET STUDY
The New York City Commercial Market (CM) Seafood Study was undertaken to measure
mercury (Hg) concentration in composite samples from seafood species most commonly
consumed by New York City residents as represented by specimens obtained from a commercial
market. Polychlorinated biphenyl (PCB) measurements were obtained in a limited subset of the
study population. Each composite sample was formed by mixing tissue from a number of
individual fish specimens into a combined amalgamated sample. The formation of the composite
sample was, in effect, a physical averaging of the individual tissue samples and the result of a
measurement on the composite sample was an estimate of the average of the specimens in the
sample. Composite sample analysis is a well established mechanism for cost effective estimation
of means of environmental samples that has a history of use in fish tissue analysis (e.g., Fabrizio
et al., 1995; Gilbert, 1987). The U.S. Environmental Protection Agency (EPA) developed a list
of the most popular species using regional and national landings, net local and national
imports/exports, domestic aquaculture production, nearby surveys of seafood species sold in
supermarkets and seafood stores, and the listing of seafood species available for sale by
individual CM wholesalers (U.S. EPA, 2008). The New Fulton Fish Market (FFM) was chosen
as the site for sample collection because it receives fish from all over the world and is the largest
seafood distributor to retailers in the United States.
The CM study is one of two complimentary studies conducted as part of an effort to
understand and respond to the results of a Health and Nutritional Examination Survey (HANES)
conducted by the New York City Department of Health and Mental Hygiene (NYCDOHMH).
The NYC HANES included measurements of blood mercury concentration in a probability
sample of 1,811 New Yorkers selected to represent the age, gender, and ethnic composition of
the adult population (McKelvey et al., 2007). The geometric mean blood mercury
(approximately equal to the median) concentration was elevated threefold compared to national
estimates. Asians registered unusually high blood mercury, with Chinese New Yorkers
registering a geometric mean almost three times that of the overall sample value. Seventy-two
percent of Chinese New Yorkers in the NYC HANES had blood Hg attaining the New York
State reportable level of 5 jig/L or above, although this is based on a small sub-sample.
Citywide, NYC HANES estimated that 1.4 million NYC adults have blood Hg at or above the
reportable level. The NYC HANES survey recorded data on the number of meals that included
fish but not the amount or type offish consumed.
1.1. SAMPLING METHODS
This section briefly describes the sample collection, compositing, and analytic methods
used in this study. For a detailed description of the procedures and protocols, you may refer to
the Quality Assurance Project Plan for the Sample Collection, Composite, and Analysis for a
Study of Mercury andPCBs in Seafood from the New Fulton Fish Market (May 2008).
-------
A ranked listing of the most commonly consumed fish species in the NYC metro area
was developed by evaluating information from national, regional, and local databases and
information sources. Databases of fishery imports, landings, and aquaculture were used to create
a ranked listing of species availability, by weight, for the NYC metro area. A regional survey of
fish availability in supermarkets, the listing offish species available for sale at the Fulton Fish
Market, and a review of the Mercury Report to Congress were also reviewed to refine the list.
The ranked listing was developed by summing together, by species, the weight of
national and local (NYC area) net edible fishery imports. Net imported fish weights were
summed together for all forms and cuts of each fish species (i.e., fresh, frozen, dried, smoked,
and pickled varieties of whole, gutted, and filleted forms offish). The latest edible fish import
information was obtained from the National Marine Fisheries Service
(http://www.st.nmfs.noaa.gov/stl/trade/index.html) and covered the period January-October
2007.
To this initial ranking, weights of National and Regional (Northeast and Middle Atlantic)
commercial fish landings were added. Annual landings for the latest commercial fish catches
(2006) were obtained from the National Marine Fisheries Service
(http://www.st.nmfs.noaa.gov/stl/commercial/landings/annual landings.html). Weights of
domestically produced food sized aquaculture species were obtained for 2005 (the latest year
available) from the U.S. Department of Agriculture's Census of Aquaculture report entitled
"Summary of Aquaculture Products Sold by Species and Size Category, United States: 2005"
(http://www.agcensus.usda.gov/Publications/2002/Aquaculture/index.asp). Weights for fish
species from the various data sources were summed together to produce the ranked listing offish
availability in the NYC Metro area.
An adjustment to, and confirmation of, the ranked listing was made by consulting
additional local and regional sources. In the report "Fish Availability in Supermarkets and Fish
Markets in New Jersey" (Burger et al., 2004), Bluefish was identified as being present in 82.5%
of markets surveyed. Bluefish was initially listed as number 45 in the ranked listing of
commonly consumed fish, but due to its popularity in NJ supermarkets it was moved up in the
rankings and included in the top 30 targeted fish species. Table 4-45 titled "Regional Popularity
of Fish and Shellfish Species - East Coast" and Table 4-46 titled "Popularity of Fish/Shellfish
Species in Restaurants - By Region - North East" in EPA's "Mercury Report to Congress,
Volume IV: An Assessment of Exposure to Mercury in the United States"
http://www.epa.gov/ttn/oarpg/t3/reports/volume4.pdf) were reviewed and confirmed the
inclusion of the fish species identified in the report were included in the final ranked list.
Lastly, the fish species listed for sale by each vendor at the New Fulton Fish Market
(FFM) in its "Seafood Products Matrix"
(http://www.newfultonfishmarket.com/products_sold.html) was reviewed to confirm that, (1) the
most frequently listed species available for sale at the FFM were included in the ranked listing,
and (2) that the top 20-30 species in the ranked listing were available for purchase at the FFM.
Minor adjustments were made to the ranked listing based on fish availability at the FFM and
-------
conversations with fish wholesalers. For example, herring was identified as being commonly
consumed in the NYC Metro area and sold by nine FFM vendors, yet only one vendor at the
FFM had herring available for sale on our sampling day. The vendor indicated other distribution
channels are used for herring, including the distribution of pickled herring directly from food
manufacturers to supermarkets. Therefore, to replace herring, the next most popular fish species
was selected for sampling.
Fresh and frozen samples were collected from a variety of wholesalers, and information
about the water body of origin was noted where available. Species were identified by wholesaler
and species name at the market and were later subjected to DNA analysis (described in Appendix
B) and visual inspection to determine the FDA approved market name (market name), common
scientific name (common name), and scientific name (i.e., genus and species). Approved market
names for seafood sold in Interstate Commerce are identified in "The Seafood List", a guidance
document and database accessible at http://www.cfsan.fda.gov/~comm/seaguid7.html that is
maintained by the FDA. Some market names included more than one species. The goal was to
obtain the required volume needed to do the analyses. The field team targeted collecting three or
more individual specimens similar in size per target species from each target wholesaler, for a
total of 45 specimens collected per species to yield 15 composite samples per species. More than
20 species were targeted. Actual sample numbers were based on practical availability at the
market.
The total length, caudal length, and weight of each specimen were recorded, and
composite samples were then created by combining multiple organisms of the same species,
roughly equal in size, and from the same wholesaler. Fish composites were prepared from edible
fillet portions with removal of skin and any surficial connective tissue or mucus. Shellfish
species included only edible tissue (all crab samples include the hepatopancreas, softshell crabs
were processed intact while hardshell crabs were shelled prior to processing, and lobster included
only muscle tissue). For fish, these composites typically included three individual specimens,
while for small shellfish (e.g., clams, oysters) the composites contained up to 200 individual
organisms. Occasionally, a composite sample for fish was comprised of two rather than three
individual specimens. This was most notable for tuna where 7 of the 15 composite samples
contained two specimens. This was due to the large size of the fish and the consequent limited
number of individual specimens at a particular vendor. The composite sampling methodology
employed provides larger sample that in turn improves the capability of the analytical methods to
achieve the target reporting limits and thus better represent fish tissue concentrations.
An analysis sample was then drawn from each composite sample and tested for total Hg
using EPA Method 245.1 Revision 3 (cold vapor technique). To help quantify sources of
variability, analyses were performed multiple times on a subset of the composite samples in two
different ways. In one case, "duplicate" analysis samples were drawn from the same composite
sample and analyzed separately. In the other case, the same analysis sample was analyzed using
the laboratory testing procedure two or three times ("replicates"). The laboratory reported an
analysis sample-specific reporting limit and a Hg concentration if Hg was detected in the assay.
-------
All total Hg concentrations for CM composite samples are reported on a wet-weight
basis. The Hg concentrations generally are reported as mg[Hg]/kg[fresh tissue weight] or parts
per million (ppm).
1.2. ORGANIZATION OF THIS REPORT
The remainder of this document describes statistical analyses of the distributions of Hg,
as measured in composite samples, for each fish and shellfish market name and species. Section
2 provides statistical analyses by market name, which is the name by which consumers typically
purchase the fish. In Section 3, the data are examined in more detail to look for trends in Hg
concentration by species (presented by "common name"), water body of origin or type of origin,
length, caudal length, and weight. Section 4 reports the variance in the data that is attributable to
measurement errors based on separate analyses of duplicate and replicate analytic results. The
Hg statistics then are placed within a risk framework (subject to caveats discussed in the text) by
comparing the mean and distribution of measured Hg concentrations with selected state and
federal action levels and by deriving estimated number of servings per week for each species that
should ensure that a woman of child-bearing age receives an average daily dose below the EPA
RfD for MeHg (see Section 5). The results also are compared to FDA monitoring data to
determine differences in the overall trends (see Section 6). Finally, the quality assurance/quality
control (QA/QC) measures that were undertaken in the analysis of the data are detailed in
Section 7. Appendix A provides the detailed statistical results of the Hg concentration analysis
and Appendix B describes the DNA analysis performed on a subset of the samples. A
description of the statistical analysis of PCB concentrations is provided in Appendix C as a
separate set of analyses of the CM fish tissue. A table of the calculated number of servings is
provided in Appendix D and the comments from independent peer reviewers are in Appendix E.
-------
2. RESULTS: MERCURY CONCENTRATIONS ACROSS SPECIES
To understand Hg concentration trends in the CM data, a number of statistical analyses
were conducted by market name. The market name was selected as the unit of analysis because
it is the name under which consumers generally identify and purchase fish and shellfish. A
crosswalk is presented in Section 2.1 (see Table 1) to link market names to species, with water
body of origin indicated, when possible, to differentiate species from the Atlantic and Pacific
Oceans. Then, Section 2.2 presents the statistical analyses by market name.
2.1. MARKET, COMMON, AND SCIENTIFIC NAMES
Table 1 presents a crosswalk table linking the market names analyzed in the CM study to
the scientific name of the one or more species encompassed by the market name. The table is
organized taxonomically and also presents one or more common names for each species. The
ocean of origin (i.e., Pacific or Atlantic if applicable) also is presented based on the vendor-
stated water body of origin for each composite sample from the CM.
Of the 33 market names, 23 referred to a single species. The remaining 10 market names
encompassed more than one species: snapper (five different species), flounder/fluke/sole (four
different species.), oyster (two different species), squid (three different species), catfish (three
different species), shrimp (two different species), whiting (two different species), cod (two
different species), bass (two different species), and tuna (two different species). The fish
samples collected were split between both Atlantic and Pacific Ocean origins, and included some
farmed species.
-------
Table 1. Crosswalk of market names to scientific names
Order
Family
Market
Name
Common Name
Genus
Species
Predominant
Location
Phylum Mollusca, Class Bivalvia (bivalves)
Ostreoida
Ostreoida
Mytiloida
Veneroida
Pectinidae
Ostreidae
Mytilidae
Veneridae
Scallop
Oyster
Mussel
Clam
Sea Scallop
Eastern Oyster
Pacific Oyster
Blue Mussel
Hardshell Clam/
Quahog/ Northern Clam /
Little Neck Clam /
Cherry Stone
Placopecten
Crassostrea
Crassostrea
Mytilus
Mercenaria
magellanicus
virginica
gigas
edulis
mercenaria
Atlantic
Atlantic coast
Pacific
Atlantic coast
Atlantic coast
Phylum Mollusca, Class Cephalopoda, Subclass Coleoidea, Superorder Decabrachia
Teuthida
Teuthida
Loliginidae
Ommastrephida
Squid
Squid
Longfm Squid
Japanese Flying Squid
Illex
Loligo
Todarodes
spp.
pealeii
pacificus
Pacific
Atlantic
Pacific
Phylum Arthropoda, Subphylum Crustacea, Class Malacostraca, Subclass Eumlacostraca, Superorder Eucarida
Decapoda
Decapoda
Decapoda
Penaeidae
Homaridae
Portunidae
Shrimp
Lobster
Blue Crab
White Shrimp
Black Tiger Shrimp
American Lobster
Blue Crab; hard and
softshell
Litopenaeus
Penaeus
Homarus
Callinectes
vannamei
monodon
americanus
sapidus
Farmed Ecuador,
Southeast Asia, and
Central America
Farmed India
Atlantic
Atlantic
Phylum Chordata, Subphylum Vertebrata, Class Chondrichthyes, Subclass Elasmobranchii, Superorder Euselachii
Rajiformes
Rajidae
Skate
Winter Skate
Leucoraja
ocellata
Atlantic
-------
Table 1. Crosswalk of market names to scientific names (continued)
Order
Family
Market Name
Common Name
Genus
Species
Predominant
Location
Phylum Chordata, Subphylum Vertebrata, Superclass Osteichthyes, Class Actinopterygii, Subclass Neopterygii (rayed fish)
Clupeiformes
Gadiformes
Gadiformes
Gadiformes
Lophiiformes
Siluriformes
Clupeidae
Gadidae
Merlucciidae
Gadidae
Lophiidae
Ictaluidae
Herring
Pollock
Whiting
Cod
Monkfish
Catfish
Atlantic Herring
Pollock
Offshore Whiting
Silver Hake
Atlantic Cod
Pacific Cod
Monkfish/Goose-
fish/Anglerfish
White catfish
Blue catfish
Channel catfish
Clupea
Pollachius
Merluccius
Merluccius
Gadus
Gadus
Lophius
Ameiurus
Ictalurus
Ictalurus
harengus
virens
albidus
bilinearis
morhua
macrocephalus
americanus
catus
furcatus
punctatus
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Pacific
Atlantic
Atlantic
Atlantic
Atlantic
Phylum Chordata, Subphylum Vertebrata, Superclass Osteichthyes, Class Actinopterygii, Subclass Neopterygii (rayed fish)
Perciformes
Perciformes
Perciformes
Percifonnes
Percifonnes
Cichlidae
Coryphaenidae
Lutjanidae
Moronidae
Nototheniidae
Tilapia
Mahi-mahi
Snapper
Bass
Chilean Sea
Bass
Tilapia
Mahi-mahi/ Dolphin Fish
Red Snapper
Caribbean Red Snapper
Lane Snapper
Yellowtail Snapper
Vermilion Snapper
Hybrid Striped Bass
Striped Bass
Chilean Sea Bass
Oreochromis
Coryphaena
Lutjanus
Lutjanus
Lutjanus
Ocyurus
Rhomboplites
Morone
Morons
Dissostichus
spp.
hippurus
campechanus
purpureus
synagris
chrysurus
aurorubens
chrysopes x saxatilisa
saxatilis
eleginoides
Foreign farmed
Pacific
Atlantic
S Atlantic
Pacific
Atlantic
Atlantic
Atlantic
Atlantic
Pacific
-------
Table 1. Crosswalk of market names to scientific names (continued)
Order
Perciformes
Perciformes
Perciformes
Perciformes
Perciformes
Perciformes
Perciformes
Perciformes
Salmoniformes
Pleuronectiformes
Pleuronectiformes
Salmoniformes
Perciformes
Family
Pomatomidae
Sciaenidae
Scombridae
Scombridae
Scombridae
Serranidae
Serranidae [formerly
Sebastidae]
Sparidae
Salmonidae
Pleuronectidae
Pleuronectidae
Salmonidae
Xiphiidae
Market Name
Blueflsh
Croaker
Spanish
Mackerel
Mackerel
Tuna
Sea Bass
Ocean Perch
Porgy
Atlantic
Salmon
Halibut
Flounder/
Fluke/
Sole
Rainbow Trout
Swordflsh
Common Name
Bluefish
Atlantic Croaker
Spanish Mackerel
Atlantic Mackerel
Yellowfin Tuna
Bigeye Tuna
Black Sea Bass
Ocean Perch
Porgy /Scup
Atlantic Salmon
Pacific Halibut
Gray Sole
Yellowtail Flounder
Summer Flounder
Blackback Flounder
Rainbow Trout
Swordfish
Genus
Pomataomus
Micropogonias
Scomberomorus
Scomber
Thunnus
Thunnus
Centropristis
Serranus [formerly
Sebastes]
Stenotomus
Salmo
Hippoglossus
Glyptocephalus
Limanda
Paralichthys
Psuedopleuronectes
Oncorhynchus
Xiphias
Species
saltatrix
undulates
maculatus
scombrus
albacares
obesus
striata
scriba [formerly
marinus]
chrysops
salar
stenolepis
cynoglossus
ferruginea
dentatus
americanus
Mykiss
gladius
Predominant
Location
N Atlantic
N Atlantic
Atlantic
Atlantic
Pacific
Pacific
Atlantic
N Atlantic
Atlantic
Atlantic and Pacific
Pacific
Atlantic
Atlantic
Atlantic
Atlantic
Farmed Idaho
Atlantic and Pacific
"Not recognized as a valid species by the Integrated Taxonomic Information System (http://www.itis.gov/).
bDNA evidence (presented in Appendix A) strongly suggests that at least one of these samples is notDissostichus eleginoides, but instead Dissostichus mawsoni. Some sources refer to both species as
Chilean Sea Bass, but there is not a clear consensus.
oo
-------
2.2. MERCURY CONCENTRATIONS BY MARKET NAME
Table 2 presents statistical descriptors of Hg concentration by market name. This table
presents the number of composite samples analyzed for each market name and indicates the
number of non-detects and detects. A non-detect refers to a measured Hg concentration that is
below the laboratory analysis reporting limit. The percent of samples that were non-detects is
also presented in this table. The mean Hg concentration was calculated for all species with at
least two composite samples; if only one composite sample was taken, the mean is equal to the
single measurement. If there are two or more composite samples, the standard deviation,
coefficient of variation, minimum, maximum, and lower and upper confidence limits on the
mean also are presented. For market names that have 10 or more composite samples, the
empirical 25* percentile, 50* percentile (median), and 75* percentile levels are also presented.
Extreme upper and lower percentiles (e.g., the 5* and 95* percentiles) were not calculated
because of the relatively small number of samples for each market name. The maximum and
minimum concentrations are reported because they provide some information about the potential
range of Hg concentrations.
In order to estimate mean Hg concentrations and to provide statistical descriptors of the
distribution of Hg concentrations across composite samples for the species with one or more
non-detect sample measurements, an assumption must be made about the composite sample Hg
concentrations that were below the reporting limit. In the statistical analyses, non-detect Hg
concentrations were estimated in two different ways:
• A non-detect was assumed to equal one-half the reporting limit. This method assumes
the actual Hg concentration has an equal probability of taking on all values between zero
and the reporting limit, so the expected value is one-half the reporting limit. Results
presented in the main body of the text use this assumption.
• A non-detect was assumed to equal the reporting limit. This method is a conservative
assumption and will yield the highest possible Hg concentrations. Results presented in
Appendix A use this assumption.
This was done to provide a sense of the impact of these assumptions on the data analysis.
Discussions of the differences between the approaches are provided at appropriate points in the
text.
Many of the species have multiple measured Hg concentrations for the same composite
sample; these are referred to as "duplicates" or "replicates". Samples labeled as duplicates
indicate that two laboratory analysis samples were drawn from the same composite and analyzed
separately. Samples labeled as replicates indicate that multiple measurements of Hg
concentration were performed on the same analysis sample. For the purposes of statistical
analysis, the duplicates or the replicates were averaged to obtain a single Hg concentration for
the respective composite sample. In one special case (Group ID 237, market name Mackerel),
three replicates resulted in one Hg non-detect and two detected Hg measurements. This
-------
composite sample was classified as a non-detect to reflect the uncertainty in the overall average;
however, the average includes the two detected measurements, along with the assumed non-
detect value. The term "aggregated composite samples" hereafter refers to the set of
concentration measurements used in the statistical analysis in which duplicates and replicates
have been averaged together, non-detects have an assumed value, and all other composite sample
measurements are as they were reported in the raw data.
The number of aggregated composite samples varies substantially across market names.
Market names have as few as one aggregated composite sample per each market name (Chilean
sea bass, halibut, herring, lobster, mahi-mahi, ocean perch, and rainbow trout), to greater than 10
aggregated composite samples per market name (sea bass, cod, blue crab, flounder/fluke/sole,
monkfish, snapper, squid, tilapia, and tuna). It should be noted that each composite sample
contained more than one organism, so even market names with one or a few composite samples
were still effectively averaged over multiple individual organisms. Non-detects occurred more
frequently with shellfish (clam, blue crab, mussel, oyster, and scallop), although a few fish had
non-detects as well (flounder/fluke/sole, mackerel, and Atlantic salmon).
Looking at standard deviations across composite samples gave an estimate of the
variability; however, because there were multiple organisms in each composite sample, it tended
to underestimate the population standard deviation. Therefore, for this analysis the population
standard deviation was estimated. To make this calculation, it was assumed that in a given
species, each composite was made of k fish. In addition, it was assumed that the fish in each
composite were statistically independent and the weight of each individual fish used in the
sample was approximately equal. If the composites had a meanM, a standard deviation s, and a
variance of V, then the individual fish distribution (called here the "estimated population
distribution") had a mean of M and a variance of &Ffor, equivalently, a standard deviation of
k'/2s, where s was the standard deviation across composite samples and equals V/2). In the CM
sample preparation, the number of specimens in each composite in a single species was not
always the same, so k actually varied across composite samples; in these cases, the harmonic
mean of the numbers offish in each composite was used to represent k. The harmonic mean was
less influenced by large outliers (or small) than the arithmetic mean because in mathematical
terms it is a lower bound on the median of a set of values and was a better representation of the
sample size in calculation of the variance. In the tuna species, 6 of the 14 composites were
composed of two fish rather than three. The harmonic mean of the number offish in the
composite samples was then 2.47 and the standard deviation across the composites was 0.25, so
the estimated population standard deviation was (2.47)2 x 0.25 = 0.39. For swordfish, all
composite samples were composed of three fish, so the estimated population standard deviation
was (3)2 x 0.019 = 0.033. The upper and lower estimated confidence limits were then calculated
assuming a normal distribution and using a sample size of k x n, where n is the number of
composite samples.
In Table 2, species are listed in order of decreasing mean Hg concentration. Mean Hg
concentrations across composite samples for the same market name range from as low as 0.0054
10
-------
mg/kg (shrimp) to as high as 0.42 mg/kg (tuna). The species with the four highest Hg
concentrations were tuna, swordfish, Spanish mackerel, and mahi-mahi, while shellfish tended to
have the lowest concentrations. Where medians could be calculated, they were generally similar
to the means. Because medians are less influenced by outliers, similarity between means and
medians indicated that the distributions were relatively symmetric and little influenced by
outliers in a single direction (e.g., several very high or very low measurements).
The results in Table 2 assume Hg non-detects are equivalent to one-half the reporting
limit. To provide a sense of the impact this assumption on the analysis, the data were also
analyzed assuming the Hg concentration in samples analyzed as non-detects were equal to the
reporting limit and these results are presented in Table A-l in Appendix A. The results for any
market names without non-detects are the same in both tables. For market names that included
non-detect concentrations, the percent difference in the mean concentration between tables
ranged from 4% for whiting to 72% for shrimp. However, the largest percent difference in mean
Hg concentrations between Table 2 (non-detects equal to one-half reporting limit) and Table A-2
(non-detects equal to reporting limit) occurred in species with low Hg concentrations, as
expected. The highest absolute change in the mean between Tables 2 and A-2 was only 0.0039
mg[Hg]/kg[fish wet weight].
11
-------
Table 2. Statistical information by market name, non-detects equal to half the reporting limit
a,b
Market Name of
Species
Tuna
Swordfish
Mahi-mahi
Spanish
Mackerel
Halibut
Bluefish
Chilean Sea
Bass
Pollock
Monkfish
Porgy
Croaker
Sea Bass
Lobster
Skate
Flounder /
Fluke / Sole
Snapper
Catfish
Number of Composite
Samples0
Del.
14
4
1
3
1
3
1
9
10
6
9
11
1
13
13
16
7
N.D.
0
0
0
0
0
0
0
0
0
0
0
0
0
1
2
0
0
Total
14
4
1
3
1
3
1
9
10
6
9
11
1
14
15
16
7
Perc.
N.D.
(%)
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
7%
13%
0%
0%
Mean
(mg/kg)
0.42
0.4
0.22
0.15
0.15
0.15
0.13
0.13
0.11
0.098
0.084
0.075
0.069
0.06
0.051
0.049
0.044
Comp.
S.D.
(mg/kg)
0.25
0.19
N/A
0.045
N/A
0.023
N/A
0.034
0.044
0.023
0.024
0.021
N/A
0.035
0.028
0.022
0.023
Est. Pop.
S.D.
(mg/kg)
0.39
0.33
N/A
0.078
N/A
0.04
N/A
0.057
0.069
0.04
0.043
0.036
N/A
0.06
0.051
0.039
0.041
Est. Pop.
C.V. (%)
93%
82%
N/A
51%
N/A
27%
N/A
44%
65%
41%
51%
49%
N/A
99%
100%
80%
93%
Min.
(mg/kg)
0.043
0.14
N/A
0.11
N/A
0.12
N/A
0.079
0.054
0.068
0.056
0.03
N/A
0.005
0.0047
0.017
0.024
25th Perc.
(mg/kg)
0.23
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.073
N/A
N/A
0.064
N/A
0.03
0.038
0.032
N/A
Median
(mg/kg)
0.39
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.095
N/A
N/A
0.078
N/A
0.064
0.049
0.044
N/A
75th Perc.
(mg/kg)
0.57
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.14
N/A
N/A
0.087
N/A
0.081
0.066
0.068
N/A
Max.
(mg/kg)
0.82
0.57
N/A
0.2
N/A
0.16
N/A
0.18
0.18
0.13
0.13
0.11
N/A
0.12
0.1
0.083
0.094
Est.
Lower
95% C.L.
on Mean*1
0.29
0.22
N/A
0.1
N/A
0.12
N/A
0.11
0.08
0.079
0.069
0.063
N/A
0.042
0.037
0.038
0.026
Est.
Upper
95% C.L.
on Mean*1
0.55
0.59
N/A
0.2
N/A
0.17
N/A
0.15
0.13
0.12
0.1
0.088
N/A
0.078
0.065
0.06
0.061
-------
Table 2. Statistical information by market name, non-detects equal to half the reporting limit' (continued)
a,b
Market Name of
Species
Cod
Whiting
Bass
Mackerel
Ocean Perch
Herring
Oyster
Blue Crab
Tilapia
Squid
Mussel
Rainbow
Trout
Clam
Atlantic
Salmon
Scallop
Shrimp
Number of Composite
Samples0
Del.
10
7
3
7
1
1
6
8
5
10
5
1
3
3
1
1
N.D.
0
1
0
1
0
0
2
O
6
2
2
0
4
6
6
6
Total
10
8
3
8
1
1
8
11
11
12
7
1
7
9
7
7
Perc.
N.D.
(%)
0%
13%
0%
13%
0%
0%
25%
27%
55%
17%
29%
0%
57%
67%
86%
86%
Mean
(mg/kg)
0.031
0.028
0.025
0.022
0.022
0.022
0.015
0.015
0.014
0.014
0.012
0.012
0.0081
0.0081
0.0055
0.0054
Comp.
S.D.
(mg/kg)
0.012
0.021
0.019
0.0078
N/A
N/A
0.014
0.0091
0.013
0.0062
0.0057
N/A
0.0058
0.0055
0.0022
0.0025
Est. Pop.
S.D.
(mg/kg)
0.019
0.051
0.033
0.017
N/A
N/A
0.063
0.03
0.023
0.017
0.048
N/A
0.028
0.0095
0.0088
0.017
Est. Pop.
C.V. (%)
63%
180%
130%
76%
N/A
N/A
410%
200%
160%
120%
390%
N/A
350%
120%
160%
310%
Min.
(mg/kg)
0.016
0.0048
0.014
0.013
N/A
N/A
0.0046
0.0043
0.0046
0.0042
0.0044
N/A
0.0043
0.0043
0.0043
0.0042
25th Perc.
(mg/kg)
0.024
N/A
N/A
N/A
N/A
N/A
N/A
0.007
0.0049
0.011
N/A
N/A
N/A
N/A
N/A
N/A
Median
(mg/kg)
0.027
N/A
N/A
N/A
N/A
N/A
N/A
0.017
0.005
0.014
N/A
N/A
N/A
N/A
N/A
N/A
75th Perc.
(mg/kg)
0.038
N/A
N/A
N/A
N/A
N/A
N/A
0.023
0.021
0.017
N/A
N/A
N/A
N/A
N/A
N/A
Max.
(mg/kg)
0.049
0.075
0.047
0.034
N/A
N/A
0.047
0.029
0.038
0.024
0.019
N/A
0.02
0.019
0.011
0.011
Est.
Lower
95% C.L.
on Mean*1
0.024
0.013
0.0034
0.017
N/A
N/A
0.0059
0.0098
0.0069
0.01
0.0081
N/A
0.0038
0.0045
0.0038
0.0036
Est.
Upper
95% C.L.
on Mean*1
0.038
0.043
0.047
0.028
N/A
N/A
0.025
0.021
0.022
0.017
0.017
N/A
0.012
0.012
0.0071
0.0073
-------
Table 2. Statistical information by market name, non-detects equal to half the reporting limit' (continued)
a,b
Market Name of
Species
Total
Number of Composite
Samples0
Del.
194
N.D.
42
Total
236
Perc.
N.D.
(%)
18%
Mean
(mg/kg)
0.075
Comp.
S.D.
(mg/kg)
0.12
Est. Pop.
S.D.
(mg/kg)
N/A
Est. Pop.
C.V. (%)
N/A
Min.
(mg/kg)
0.0042
25th Perc.
(mg/kg)
0.014
Median
(mg/kg)
0.034
75th Perc.
(mg/kg)
0.083
Max.
(mg/kg)
0.82
Est.
Lower
95% C.L.
on Mean*1
0.059
Est.
Upper
95% C.L.
on Mean*1
0.091
aDet. is the number of samples with results above the reporting limit (referred to here as "detects"), N.D. is the number of samples with results below the reporting limit, total is the total number of samples for the
species (excluding replicate and duplicate samples), Perc. N.D. is the percent of the total number of samples which are below the reporting limit, Comp. S.D. is the standard deviation of the mean of the composite
sample Hg concentrations for the market name, Est. Pop. S.D. is the estimated population standard deviation, Est. Pop. C.V. is the estimated coefficient of variation calculated as the population standard
deviation/mean x 100%, 25th Perc. is the 25th percentile, 75th Perc. is the 75th percentile, est. lower 95% C.L. on mean is the estimated lower 95th percent confidence limit on the mean, and Est. Upper 95% C.L. on
mean is the estimated upper 95th percent confidence limit on the mean.
bNumber of Market Name groups included = 33 .
'Number of composite samples after averaging the duplicates and replicates for a single composite sample and treating the average as a single point.
^Calculated based on the estimated population standard deviation.
NA = Value not available.
-------
Composite samples taken for one market name—tuna—contained an outlier with
significantly lower Hg concentrations than in the other tuna samples. This one composite
consisted of three replicate analysis samples, with consistently low Hg concentrations. It was
impossible to determine from the data collected the reason for the lower measured concentration.
The average across these replicates is 0.0427 mg/kg, while the next highest Hg concentration for
a tuna composite sample was 0.13 mg/kg. Table 3 presents the effects on the mean, S.D., C.V.,
and confidence limits of including and excluding this outlier. Without the outlier, the mean Hg
concentration is slightly (7%) higher, and the S.D. decreases, reflecting the lower variation
across samples when the outlier is removed. Because mean tuna concentration is not strongly
sensitive to the outlier, it was not removed for all subsequent analyses.
Table 3. Examination of the effect of outlier data in the tuna species
a,b,c
Market Name of
Species
Tuna, Including
Outliers
Tuna, Excluding
Outliers
Number of Composite Samples
Del.
14
13
N.D.
0
0
Total
14
13
Perc.
N.D. (%)
0%
0%
Mean
(mg/kg)
0.42
0.45
S.D.
(mg/kg)
0.25
0.23
c.v.
(%)
59%
52%
Lower
95% C.L.
on Mean
0.29
0.32
Upper
95% C.L.
on Mean
0.55
0.57
"Column names are as defined in Table 2.
bNon-detects were assumed to equal half the reporting limit.
'The outlier composite sample replicate group had an average measured Hg concentration of 0.0427 mg/kg in fish tissue; the data are too limited
to identify a possible reason for the low value.
15
-------
3. MERCURY CONCENTRATION BY SPECIES, LOCATION, CONDITION, AND SIZE
Differences in habitat (or microhabitat) and diet among different species sold under the
same market name (e.g., red and yellowtail snapper sold as snapper) might result in significant
differences in Hg tissue residue levels among those species. Different water bodies (e.g., Pacific
versus Atlantic) can exhibit lower or higher total Hg contamination of water, sediments, and food
webs. Hg concentrations in fish and shellfish were analyzed, therefore, by species (see
Section 3.1) and by vendor-stated water-body-of-origin categories (see Section 3.2). To better
understand the role of specimen size and age on bioaccumulation of Hg, correlations of Hg
concentration and fish or shellfish body mass and length also are presented (see Section 3.3).
3.1. SPECIES
Table 4 presents the summary statistics for each market name broken down by common
name (a crosswalk linking the common name to the scientific name can be found in Section 2.1,
Table 1). The table includes the 9 of the 33 market names which included two or more species
with more than one composite sample per species. Because many species had fewer composite
samples than available for market name categories, only the mean, S.D., C.V., maximum, and
minimum Hg concentrations are reported in Table 4. The blue crab included both hardshells
(condition for most of their annual life cycle) and softshells. Softshell crabs have recently shed
their exoskeleton (i.e., have recently molted) to allow growth to a larger size.
In general, little difference was found when comparing species within a single market
name. Tests for significance (at the/? = 0.05 level) revealed no statistically significant
differences when comparing the market name mean and the means in the individual common
names. Offshore whiting Hg concentrations were higher on average than those for silver
hake/whiting by a factor of 2.6, and summer flounder Hg concentrations exceeded those for
blackback flounder by a factor of 2.4. For whiting, the organisms sampled from the CM were of
similar length and mass for both species (i.e., offshore whiting and silver hake), suggesting size
(as a proxy for age) is not responsible for this difference. For flounder/fluke/sole market name,
however, the summer flounder sampled tended to be consistently larger than the blackback
flounder from the CM. Hardshell blue crabs exhibited higher Hg concentrations than softshell
crabs by a factor of 2.1; softshell crabs had multiple non-detects, while all hardshell crab
composite samples had detectable Hg concentrations. Softshell crabs also tended to have a
significantly higher water content, which would dilute Hg concentrations, than hardshell crabs—
particularly hardshells in the few weeks prior to molting when body tissues essentially fill the
shell. The conservative assumption that the Hg concentrations in the non-detect samples were
equal the full reporting limit, as shown in Table A-3 in Appendix A, does not alter any of the
conclusions stated above.
16
-------
Table 4. Mercury concentrations by species or by condition (e.g., hard- or softshell) within a market name; non-
detects equal to half the reporting limita'b
Market Name
Catfish
Codd
Blue Crab
Flounder/ Fluke/
Sole
Species
Catfish, Blue
Catfish, Channel
Catfish, White
All Catfish
Cod, Atlantic
Cod, Pacific
All Cod
Blue Crab/Hardshell
Blue Crab/Softshell
All Crab
Flounder, Blackback
Flounder, Summer
Sole, Gray
All Flounder/Fluke/Sole
Number of Composite
Samples0
Del.
2
4
1
7
7
3
10
5
3
8
4
5
4
13
N.D.
0
0
0
0
0
0
0
0
3
3
1
0
0
1
Total
2
4
1
7
7
3
10
5
6
11
5
5
4
14
Perc.
N.D.
(%)
0%
0%
0%
0%
0%
0%
0%
0%
50%
27%
20%
0%
0%
13%
Mean
(mg/kg)
0.037
0.048
0.041
0.044
0.031
0.030
0.031
0.021
0.010
0.015
0.031
0.073
0.059
0.051
Comp. S.D.
(mg/kg)
0.012
0.032
N/A
0.023
0.013
0.009
0.012
0.003
0.010
0.009
0.017
0.026
0.012
0.028
Est. Pop. S.D.
(mg/kg)
0.021
0.055
N/A
0.041
0.022
0.015
0.019
0.013
0.028
0.030
0.032
0.045
0.020
0.051
Est. Pop.
C.V. (%)
57%
110%
N/A
93%
70%
49%
63%
62%
270%
200%
100%
61%
34%
100%
Min.
(mg/kg)
0.028
0.024
N/A
0.024
0.016
0.024
0.016
0.017
0.004
0.004
0.005
0.045
0.044
0.005
Max.
(mg/kg)
0.045
0.094
N/A
0.094
0.049
0.040
0.049
0.025
0.029
0.029
0.049
0.100
0.071
0.100
-------
Table 4. Mercury concentrations by species or by condition (e.g., hard- or softshell) within a market name; non-
detects equal to half the reporting limita'b (continued)
Market Name
Shrimp
Snapper
Squid d
Tuna
Species
Shrimp, Black Tiger
Shrimp, White
All Shrimp
Snapper, Caribbean Red
Snapper, Lane
Snapper, Red
Snapper, Vermilion
Snapper, Yellowtail
All Snapper
Squid, Japanese Flying
Squid, Longfin (Atlantic)
All Squid
Yellowfin Tuna
Bigeye Tuna
All Tuna
Number of Composite
Samples0
Del.
0
1
1
1
2
6
3
4
16
3
7
10
7
7
14
N.D.
2
4
6
0
0
0
0
0
0
1
0
1
0
0
0
Total
2
5
7
1
2
6
O
4
16
4
7
11
7
7
14
Perc.
N.D.
(%)
100%
80%
86%
0%
0%
0%
0%
0%
0%
25%
0%
17%
0%
0%
0%
Mean
(mg/kg)
0.004
0.006
0.005
0.039
0.040
0.057
0.040
0.051
0.049
0.010
0.017
0.014
0.420
0.410
0.420
Comp. S.D.
(mg/kg)
0.000
0.003
0.003
N/A
0.018
0.020
0.037
0.022
0.022
0.004
0.005
0.006
0.280
0.230
0.250
Est. Pop. S.D.
(mg/kg)
0.001
0.020
0.017
N/A
0.031
0.035
0.065
0.038
0.039
0.010
0.017
0.017
0.440
0.360
0.390
Est. Pop.
C.V. (%)
13%
340%
310%
N/A
78%
63%
160%
76%
80%
92%
100%
120%
100%
87%
93%
Min.
(mg/kg)
0.004
0.004
0.004
N/A
0.027
0.031
0.017
0.034
0.017
0.005
0.012
0.004
0.043
0.130
0.043
Max.
(mg/kg)
0.004
0.011
0.011
N/A
0.052
0.076
0.083
0.082
0.083
0.015
0.024
0.024
0.800
0.820
0.820
oo
-------
Table 4. Mercury concentrations by species or by condition (e.g., hard- or softshell) within a market name; non-
detects equal to half the reporting limita'b (continued)
Market Name
Whiting
Species
Whiting, Offshore
Whiting/Silver Hake
All Whiting
Number of Composite
Samples0
Del.
2
5
7
N.D.
0
1
1
Total
2
6
8
Perc.
N.D.
(%)
0%
17%
13%
Mean
(mg/kg)
0.052
0.020
0.028
Comp. S.D.
(mg/kg)
0.032
0.011
0.021
Est. Pop. S.D.
(mg/kg)
0.062
0.028
0.051
Est. Pop.
C.V. (%)
120%
140%
180%
Min.
(mg/kg)
0.029
0.005
0.005
Max.
(mg/kg)
0.075
0.033
0.075
"Column names are as defined in Table 2.
bNumber of Market Names encompassing multiple species or conditions equals 10 of the 33 total Market Names.
'Number of composite samples after averaging the duplicates and replicates for a single composite sample and treating the average as a single point.
''Species from different oceans.
NA = Value not available because only one composite sample was analyzed.
-------
3.2. WATER BODY OF ORIGIN
Water body (e.g., Atlantic or Pacific) of origin and source type (e.g., farmed or wild)
were used to determine if there were any differences in Hg concentrations within a species or
among species for a given market name that could be attributed to the water body of origin or the
source type for the fish marketed at the CM. Vendor-identified water bodies of origin (included
in the CM database) were used to assign composite samples to the following uniform categories:
"Atlantic", "Pacific", "Bay", "Sound", "Lake", "River", "Foreign Farmed", and "Unknown". In
addition, Atlantic and Pacific water bodies were also further separated into "Wild" and "Farmed"
categories. All market names which had at least two composite samples from at least two
different water bodies were included in the analysis.
Table 5 shows the seven (out of the 33) market names that included at least two
composite samples originating from at least two locations. Because the water body
subcategories have fewer composite samples than the market names categories, only mean, S.D.,
C.V., maximum, and minimum values are reported in Table 5.
Tests for significance (at the/? = 0.05 level) revealed no statistically significant
differences when comparing the market name mean and the means in the individual water bodies
of origin. No clear trend was found for concentrations of Hg in organisms caught in the Atlantic
versus Pacific Oceans. Wild cod caught from both the Atlantic and Pacific oceans had nearly
identical mean Hg concentrations. Longfm squid from the Atlantic Ocean had nearly double the
Hg concentration of Japanese flying squid from the Pacific Ocean. Swordfish caught in the
Pacific had approximately 50% more Hg than those from the Atlantic Ocean; data on the original
size of the fish from which swordfish steaks were cut were inadequate to determine if the effect
was due to differences in the average size offish harvested from each ocean. Note that at this
level of categorization, the sample sizes for each group were very small. When the non-detects
were assumed to equal the full reporting limit, as shown in Table A-4 of Appendix A, none of
the conclusions above were altered.
20
-------
Table 5. Mercury concentrations by water body of origin within a market name species; non-detects equal to
half the reporting limit3
Market
Name of
Species
Clam
Cod
Mussel
Salmon
Snapper
Water Body or Water Type
of Origin
Clam, Atlantic, Wild
Clam, Farmed
Clam, Long Island Sound
All Clam
Cod, Atlantic, Wild
Cod, Pacific, Wild
All Cod
Mussel, Atlantic, Wild
Mussel, Farmed
All Mussel
Atlantic Salmon, Wild
Atlantic Salmon, Farmed
All Salmon
Snapper, Atlantic, Wild
Snapper, Pacific, Wild
Snapper, Unknown
All Snapper
Number of Composite
Samples'"
Del.
1
2
0
O
7
O
10
4
1
5
0
O
O
13
2
1
16
N.D.
3
0
1
4
0
0
0
1
1
2
3
3
6
0
0
0
0
Total
4
2
1
7
7
O
10
5
2
7
O
6
9
13
2
1
16
Perc.
N.D.
(%)
75%
0%
100%
57%
0%
0%
0%
20%
50%
29%
100%
50%
67%
0%
0%
0%
0%
Mean
(mg/kg)
0.0082
0.0095
0.0047
0.0081
0.031
0.03
0.031
0.013
0.0097
0.012
0.0044
0.0099
0.0081
0.05
0.04
0.047
0.049
Comp. S.D.
(mg/kg)
0.0079
0.00071
N/A
0.0058
0.013
0.0086
0.012
0.0055
0.0075
0.0057
0.00022
0.0061
0.0055
0.024
0.018
N/A
0.022
Est. Pop.
S.D.
(mg/kg)
0.037
0.0041
N/A
0.028
0.022
0.015
0.019
0.046
0.064
0.048
0.00038
0.01
0.0095
0.042
0.031
N/A
0.039
Est. Pop.
C.V. (%)
450.0%
43.0%
N/A
350.0%
70.0%
49.0%
63.0%
340.0%
660.0%
390.0%
8.5%
110.0%
120.0%
83.0%
78.0%
N/A
80.0%
Min.
(mg/kg)
0.0043
0.009
N/A
0.0043
0.016
0.024
0.016
0.0044
0.0044
0.0044
0.0043
0.0044
0.0043
0.017
0.027
N/A
0.017
Max.
(mg/kg)
0.02
0.01
N/A
0.02
0.049
0.04
0.049
0.019
0.015
0.019
0.0047
0.019
0.019
0.083
0.052
N/A
0.083
-------
Table 5. Mercury concentrations by water body of origin within a market name species; non-detects equal
to half the reporting limit" (continued)
Market
Name of
Species
Squid
Swordfish
Water Body or Water Type
of Origin
Squid, Atlantic, Wild
Squid, Pacific, Wild
All Squid
Swordfish, Atlantic, Wild
Swordfish, Pacific, Wild
All Swordfish
Number of Composite
Samples'"
Del.
7
3
10
2
2
4
N.D.
0
2
2
0
0
0
Total
7
5
12
2
2
4
Perc.
N.D.
(%)
0%
40%
17%
0%
0%
0%
Mean
(mg/kg)
0.017
0.0092
0.014
0.33
0.48
0.4
Comp. S.D.
(mg/kg)
0.0048
0.0046
0.0062
0.26
0.13
0.19
Est. Pop.
S.D.
(mg/kg)
0.017
0.0098
0.017
0.45
0.22
0.33
Est. Pop.
C.V. (%)
100.0%
110.0%
120.0%
140.0%
46.0%
82.0%
Min.
(mg/kg)
0.012
0.0042
0.0042
0.14
0.39
0.14
Max.
(mg/kg)
0.024
0.015
0.024
0.51
0.57
0.57
to
to
"Column names are as defined in Table 2.
bNumber of composite samples after averaging the duplicates and replicates for a single composite sample and treating the average as a single point.
NA = Value not available because only one composite sample was analyzed.
-------
3.3. EFFECTS OF BODY WEIGHT AND LENGTH ON MERCURY
CONCENTRATIONS
MeHg in aquatic food webs can bioaccumulate in the consumer organisms over time,
suggesting that older fish will tend to have higher tissue MeHg and total Hg concentrations.
Both the length and weight of many invertebrates and fish within a single species can act as a
proxy for age; however, depending on food availability and water temperature, individuals of the
same species may grow at substantially different rates. We examined correlations between total
Hg concentrations and body weight, length, and caudal length (fish-only measure)1 to determine
whether significant trends between measures of size and Hg concentrations were present in the
data. Only fish weights and lengths that were reported for the entire fish were used in the
analysis, and not all the market names could be analyzed since only a fillet or some other partial
sample (e.g., steak) was collected from the CM. For fish species, the different lengths and
weights for each of the fish in a composite sample were averaged to get the average length and
weight for that sample. For shellfish, the total composite weight was divided by the number of
organisms in the composite to get the average weight. Shrimp were included in the analysis,
although they had been headed prior to collection.
9
The squares of the correlation coefficients (R ) between Hg concentrations and the
length, caudal length, and weight were calculated, and graphs were generated to visualize the
relationships by species. Generally, concentration tends to increase as measures of length and
r\
weight increase. However, the correlations and, in turn, R values can be very sensitive to
patterns in the data that may result in improperly inflated values and findings of statistical
significance that are not meaningful. There were a number of examples of these sorts of issues
in the data. In a number of cases, large R2 values were clearly the result of large separations
between data points and/or clusters of data points. In some cases, clusters of species specific
data displayed a negative relationship between concentration and size but the overall relationship
between clusters was positive. Although the data generally show positive relationships between
measures of size and Hg concentration, the results of the tests of significance of the correlations
between size and concentration and the graphs of the data were not included because of the many
problematic cases.
3.4. WILD VS FARMED RESULTS
Only three species (clams, mussels, and salmon) had results on wild versus farmed
samples, precluding a thorough statistical analysis. However, a descriptive qualitative analysis is
provided herein. Three of the four samples of wild Atlantic clams had non-detectable
concentrations of Hg. The two farmed samples of clams had detectable concentrations of Hg;
however the concentration was low for both samples. Four of the five samples of wild Atlantic
1 The three most common measures offish length are "standard length" (from tip of the longest jaw or snout to the base of tail
fin, specifically, to the end of the hypural bone or caudal peduncle), "total length" (snout to tip of longest tail fin when lobes are
pinched together), and "fork length" (snout to the center of the fork in the caudal fin). The lengths of whole fish from the CM
were reported as "length", equivalent to total length, and "caudal length", equivalent to the definition of standard length above.
23
-------
mussels had detectable concentrations of Hg; however, the concentrations were low for all
samples. One of the two farmed samples had a detectable concentration of Hg, but the
concentration was low. All three samples of wild Atlantic salmon had non-detectable
concentrations for Hg. Three of the six samples of farmed Atlantic salmon had detectable
concentrations of Hg however, the concentrations were low.
24
-------
4. ANALYSIS OF MEASUREMENT VARIABILITY
For the analysis of the CM samples, multiple organisms were combined into a composite
sample. Next, an analysis sample was drawn from the composite and the laboratory testing
procedure was applied to measure the total Hg concentration. Since the composite sample
homogenate was not perfectly uniform, there was some variability associated with taking an
aliquot of the composite sample for preparation and analysis. In addition, the sample preparation
and analysis may have been introduced of variability, such as inherent measurement error or
calibration drift.
To help quantify both sources of variability, analyses were performed multiple times on a
subset of the composite samples. "Duplicate" analysis samples were drawn from the same
composite sample and analyzed separately. In other cases, the same analysis sample was
analyzed using the laboratory testing procedure two or three times ("replicates"). In each of
these cases, the collection of measurements within a duplicate pair or within a replicate group
was referred to as a "duplicate/replicate group". Within the data, there were 24 duplicate groups
and 15 replicate groups, and, in general, these groups were for different market name species.
In order to determine the contribution of the measurement variability to the overall Hg
variability, the variance within each duplicate and replicate group was calculated as follows:
. (*i - x)2 + (x2 -x)2 +...+ (xn - x)2
Variance = — — — —
n-l
where:
xn = Measured concentration for measurement n,
x = Mean of measured concentrations for the replicate/duplicate group, and
n = Number of measurements per replicate/duplicate group.
The variance equals the square of the S.D. within the replicate/duplicate group. Table 6
and Table 7 show the variances for the duplicates and replicates, respectively. These variances
were calculated using both the raw Hg concentrations and the natural logarithm of the Hg
concentrations assuming the non-detects were either all equal to half the reporting limit or all
equal to the reporting limit.
Because the duplicate and replicate groups were across many different market name
species, it was necessary to confirm that the variances in the measurements did not vary
significantly across the different replicate and duplicate groups before performing the full
variability analysis. The log-normal distribution is likely to give a better fit than the normal
distribution because the data showed evidence of a skew in the positive direction. The Bartlett
r\
homogeneity of variance test requires normality to be a valid test; thus, it was applied separately
2 The Bartlett homogeneity of variance test assumes the individual concentrations or log of concentrations are normally distributed. Non-detects
were replaced by half the reporting limit for this test.
25
-------
to the duplicate and replicate groups after transforming the concentrations using the natural
logarithm.
For the duplicates, the Bartlett test indicated the differences in the variances were not
significant in log space (p-value 0.87). For the replicates, the analysis found significant non-
homogeneity in the natural logarithm of the concentrations. However, an inspection of the data
revealed that the highest and lowest variances were for samples 001.C and 237.C, each of which
include non-detects. Because the true values of the non-detects were uncertain, and hence the
variances are uncertain, these two replicate groups were excluded from the analysis. After
removing these measurements, the Bartlett test was not significant in log space (p-value 0.77).
Thus, the analysis of measurement variability for replicates excluded sample groups 001 and
237.
To analyze the variability, the contributions of the measurement variability to the overall
variance amongst all the concentrations in the duplicate and replicate groups were separately
calculated. Suppose that the rth duplicate or replicate group has nt measurements xtj, j = 1, 2, ...
«;. The total sum of squares (TSS) can be written as:
where:
Grand mean = x = X X xv• / X ni
This total sum of squares can then be broken into the sum of squares between groups
plus the sum of the squares within groups (SSwithin)'-
The sum of squares within groups accounts for variation within a single group and is given by
the formula:
' J
•>
where J. denotes the mean of the nt measurements on the /'th duplicate or replicate group:
i
The sum of squares between groups accounts for the variation across groups and is given by the
formula:
SSbetween =
The degrees of freedom (df) is defined as:
26
-------
so that the variance within groups is defined as:
Variance within groups = SSwithin I df
and the S.D. within groups is the square root of the variance within the group. Finally, the
proportion of the variance explained by the variability within the duplicates or replicates is
defined as:
n /-IT • Variance within groups
Prop oj Variance =
N-\)
where N is the total number of measurements across all groups.3
Table 8 and Table 9 show the variability analysis for the duplicate groups and replicate
groups (respectively) using both the raw and log concentrations and assuming non-detects are
either all equal to half the reporting limit or all equal to the reporting limit. The calculations in
log space are shown in bold since the data are positively skewed and the lognormal distribution
is likely a better characterization of their distribution. The analysis indicates that the proportion
of the variance of a sample explained by the variation among the duplicates or replicates of that
sample is a very small (less than 1%) portion of the overall variance across all measurements in
all groups. In part, this result stems from the fact that many different species are included in the
analysis and the variance in concentration (and log concentration) is large across the groups from
different species.
To determine the effect of limiting the analysis to a single species, the calculations were
repeated for the tuna market name species which included four replicate groups and three
duplicate groups. Table 10 shows the variability analysis for the tuna duplicate and replicate
groups in both raw and log space, with the log space calculations shown in bold. In this case, the
proportion of the variance explained by the measurement variability in the duplicates remains
close to 1% (1.2%) for the log space calculations. For the replicate groups, the proportion of the
variance explained by the measurement variability was still below 1%; however, this analysis
includes the replicate group identified as an outlier in Section 2. When this replicate group was
excluded, the proportion of the variance explained by the measurement variability was again
below 1% (0.22%, not shown). This analysis suggests that even restricting the analysis to a
single species, the variability associated with the composite sampling procedure and the
3 This proportion of variance equation is an approximation to a much more complicated component of variance calculation. Similar results can be
obtained using the sum of squares within groups divided by the total sum of squares, which gives the proportion of the sum of squares explained.
27
-------
laboratory testing procedure was small. In fact, the variability was minimal and demonstrated
the high level of quality control that was achieved in generating the data.
28
-------
Table 6. Variance within individual composite sample duplicates"
Composite
Sample
Number
4
15
26
37
48
59
70
81
92
103
114
125
136
146
158
169
180
191
Number of
Duplicates
2
2
2
2
2
2
2
2
2
2
2
2
2
4
2
2
2
2
Market Name
Cod
Shrimp
Monkfish
Tuna
Flounder/Fluke/Sole
Atlantic Salmon
Monkfish
Blue Crab
Bluefish
Flounder/Fluke/Sole
Catfish
Porgy
Flounder/Fluke/Sole
Lobster
Tuna
Blue Crab
Whiting
Mussel
Non-Detects Equal to Half the Reporting Limit
Mean
2.6 x ID'02
4.3 x lO'03
1.5 x IQ-01
4.1 x 10'01
2.6 x 10-02
4.7 x 10'03
1.5 x 10'01
2.2 x 10-02
1.6 xlQ-01
7.1 x 10'02
4.5 x 10-02
8.6 x lO'02
4.5 x 10'02
7.3 x 10-02
5.0 x lO'01
2.5 x 10'02
7.5 x ID'02
1.6xlO-°2
Variance Using
Concentration
4.5 x 10'06
1.1 x ID'08
5.0 x 10'05
0.0
0.0
1.1 x ID'08
0.0
2.0 x 10'06
0.0
8.0 x 10'06
0.0
8.5 x lO'05
5.0 x 10'07
1.8x 10'05
5.0 x 10'05
5.0 x 10'07
1.3 x ID'05
0.0
Variance Using
Log of
Concentration
6.9 x 10'03
6.0 x lO'04
2.4 x 10'03
0.0
0.0
5.1 x ID'04
0.0
4.1 x 10'03
0.0
1.6 x 10'03
0.0
1.2 x ID'02
2.5 x 10'04
3.9 x 10'03
2.0 x lO'04
8.3 x 10'04
2.3 x ID'03
0.0
Non-Detects Equal to the Reporting Limit
Mean
2.6 x 10'02
8.7 x 10'03
1.5 x IQ-01
4.1 x 10'01
2.6 x 10'02
9.4 x 10'03
1.5 x 10'01
2.2 x 10'02
1.6 xlQ-01
7.1 x 10'02
4.5 x 10'02
8.6 x 10'02
4.5 x 10'02
7.3 x 10'02
5.0 x 10'01
2.5 x ID'02
7.5 x 10'02
1.6X10'02
Variance Using
Concentration
4.5 x 10'06
4.5 x ID'08
5.0 x ID'05
0.0
0.0
4.5 x ID'08
0.0
2.0 x 10'06
0.0
8.0 x 10'06
0.0
8.5 x lO'05
5.0 x 10'07
1.8x 10'05
5.0 x 10'05
5.0 x 10'07
1.3 x 10'05
0.0
Variance Using
Log of
Concentration
6.9 x 10'03
6.0 x 10'04
2.4 x 10'03
0.0
0.0
5.1 x ID'04
0.0
4.1 x ID'03
0.0
1.6x 10'03
0.0
1.2 xlQ-02
2.5 x 10'04
3.9 x 10'03
2.0 x 10'04
8.3 x 10'04
2.3 x 10'03
0.0
to
VO
-------
Table 6. Variance within individual composite sample duplicates" (continued)
Composite
Sample
Number
202
213
224
235
246
257
Number of
Duplicates
2
2
2
2
2
2
Market Name
Tuna
Croaker
Cod
Mackerel
Monklish
Scallop
Minimum Variance
Maximum Variance
Non-Detects Equal to Half the Reporting Limit
Mean
6.8 x l(r01
5.8 x icr02
2.7 x 10'02
2.1 x 10-02
1.1 x lO'01
1.1 x icr02
Variance Using
Concentration
8.0 x 10'04
5.0 x ID'07
1.2 x 10'05
5.0 x 10'07
0.0
5.0 x 10'07
0.0
8.0 x lO'04
Variance Using
Log of
Concentration
1.7x 10'03
1.5 x ID'04
1.8 x 10'02
1.2 x ID'03
0.0
4.5 x 10'03
0.0
1.8 x lO'02
Non-Detects Equal to the Reporting Limit
Mean
6.8 x 10'01
5.8 x ID'02
2.7 x 10'02
2.1 x 10'02
1.1 x lO'01
1.1 x ID'02
Variance Using
Concentration
8.0 x 10'04
5.0 x 10'07
1.2 xlQ-05
5.0 x 10'07
0.0
5.0 x 10'07
0.0
8.0 x 10'04
Variance Using
Log of
Concentration
1.7x 10'03
1.5 x ID'04
1.8x 10'02
1.2 xlO'03
0.0
4.5 x 10'03
0.0
l.SxlO'02
"Duplicates refer to multiple samples drawn from the same composite.
-------
Table 7. Variance within individual composite sample replicates"
Composite
Sample
Number
001.C
021.C
041.C
058.C
076.C
094.C
lll.C
129.C
147.C
165.C
183.C
201.C
219.C
237.C
255.C
Number of
Replicates
3
3
2
3
3
3
3
3
3
3
3
3
3
3
3
Market Name
Flounder/Fluke/Sole
Squid
Oyster
Tuna
Flounder/Fluke/Sole
Ocean Perch
Tuna
Sea Bass
Lobster
Monkfish
Tuna
Tuna
Skate
Mackerel
Skate
Minimum Variance
Maximum Variance
Non-Detects Equal to Half the Reporting Limit
Mean
4.8 x 10'03
2.4 x icr02
2.0 x icr02
4.3 x lO'02
5.7 x lO'02
2.2 x lO'02
8.2 x lO'01
6.1 x it)'02
6.7 x 10-02
8.0 x 10'02
1.9 x 10'01
8.0 x 10'01
7.4 x lO'02
1.6 x ID'02
2.7 x 10-02
Variance Using
Concentration
3.3 x ID'09
4.0 x lO'06
5.0 x 10'07
2.3 x 10'06
9.3 x lO'06
2.3 x lO'06
7.0 x 10'04
4.1 x lO'05
1.4x 10'05
3.2 x ID'05
l.OxlO'04
4.3 x 10'04
9.0 x 10'05
9.4 x 10'05
4.0 x 10'06
3.3 x ID'09
7.0 x 10'04
Variance Using
Log of
Concentration
1.4x 10'04
7.0 x 10'03
1.3 x ID'03
1.3 x ID'03
2.9 x 10'03
4.6 x 10'03
l.OxlO'03
1.2 xlQ-02
3.3 x ID'03
5.0 x 10'03
2.8 x 10'03
6.8 x 10'04
1.7xlO-°2
7.8 x 10'01
5.5 x ID'03
1.4x 10'04
7.8 x 10'01
Non-Detects Equal to the Reporting Limit
Mean
9.6 x 10'03
2.4 x 10'02
2.0 x lO'02
4.3 x 10'02
5.7 x 10'02
2.2 x 10'02
8.2 x 10'01
6.1 x lO'02
6.7 x 10'02
8.0 x 10'02
1.9 x 10'01
8.0 x 10'01
7.4 x lO'02
1.7x 10'02
2.7 x 10'02
Variance Using
Concentration
1.3 x ID'08
4.0 x 10'06
5.0 x lO'07
2.3 x 10'06
9.3 x 10'06
2.3 x 10'06
7.0 x ID'04
4.1 x lO'05
1.4 x ID'05
3.2 x 10'05
1.0 x ID'04
4.3 x ID'04
9.0 x lO'05
5.1 x ID'05
4.0 x lO'06
1.3 x 10'08
7.0 x lO'04
Variance Using
Log of
Concentration
1.4x 10'04
7.0 x 10'03
1.3 x ID'03
1.3 x ID'03
2.9 x 10'03
4.6 x 10'03
l.OxlO'03
1.2 xlQ-02
3.3 x ID'03
5.0 x 10'03
2.8 x 10'03
6.8 x 10'04
1.7xlO-°2
2.4 x 10'01
5.5 x ID'03
1.4x 10'04
2.4 x 10'01
"Replicates refer to multiple laboratory analyses performed on the same composite sample.
-------
Table 8. Variability analysis of composite sample duplicates
Total Sum of Squares (TSS)
Sum of Squares Between Groups
(BSS)
Sum of Squares Within Groups
(WSS)
Total Degrees of Freedom Within
Groups
Variance Within Groups
Standard Deviation Within Groups
Proportion of Variance Explained
by Variability Within Groups
Non-Detects Equal to Half the Reporting
Limit
Using Composite
Concentrations
1.3
1.3
1.1 x icr03
26
4.2 x 1CT05
6.4 x ICT03
0.15%
Using Log of
Composite
Concentrations
SO
SO
6.8 x 10-02
26
2.6 x 10-03
5.1 x lO'02
0.16%
Non-Detects Equal to the Reporting
Limit
Using Composite
Concentrations
1.3
1.3
1.1 x ICT03
26
4.2 x ICT05
6.4 x ICT03
0.15%
Using Log of
Composite
Concentrations
69
69
6.8 x 10-02
26
2.6 x 10-03
5.1 x lO'02
0.19%
Bold indicates the calculations in log space since the data are positively skewed and the lognormal distribution is likely a better characterization
of their distribution.
32
-------
Table 9. Variability analysis of composite sample replicates"
Total Sum of Squares (TSS)
Sum of Squares Between Groups
(BSS)
Sum of Squares Within Groups
(WSS)
Total Degrees of Freedom Within
Groups
Variance Within Groups
Standard Deviation Within Groups
Proportion of Variance Explained
by Variability Within Groups
Non-Detects Equal to Half the Reporting
Limit
Using Composite
Concentrations
2.9
2.9
2.9 x lO'03
25
1.1 x 10-04
1.1 x icr02
0.15%
Using Log of
Composite
Concentrations
53
53
1.3 x 1Q-01
25
5.1 x i
-------
5. COMPARISON OF CM DATA TO RISK METRICS
5.1. FISH ADVISORIES AS A COMMUNICATOR OF RISK
Fish advisories to the public, either at the state or federal (e.g., EPA, FDA) level, provide
a narrative statement that is reasonably easy for the public to understand and to follow (i.e.,
implement). For sport or game fish, the states generally post fish advisories by species or for
specific bodies of water or regions of the state as needed. For commercially marketed fish, the
advisories usually specify several categories offish consumption frequency (e.g., two servings
per week, one serving per week, one serving per month, never consume) for specific population
subgroups (e.g., children, women of childbearing age) and which fish market names fall into
each consumption frequency category.
Hg is of particular concern as a developmental neurotoxicant; therefore, Hg fish
advisories generally target women of childbearing age, nursing infants, and young children. The
most recent United States federal fish advisory for Hg is the 2004 EPA and FDA joint
recommendation (see text box below), although FDA might still be receiving comments on its
new risk/benefit analysis4 to support an updated advisory that attempts to balance the
developmental health benefits of consuming fish against the risks from Hg consumption.
2004 EPA and FDA Advice for:
• Women Who Might Become Pregnant;
• Women Who are Pregnant;
• Nursing Mothers;
• Young Children
By following these three recommendations for selecting and eating fish or shellfish, women and young children
will receive the benefits of eating fish and shellfish and be confident that they have reduced their exposure to the
harmful effects of Hg.
1. Do not eat Shark, Swordfish, King Mackerel, or Tilefish because they contain high levels of Hg.
2. Eat up to 12 ounces (2 average meals) a week of a variety of fish and shellfish that are lower in Hg.
• Five of the most commonly eaten fish that are low in Hg are shrimp, canned light tuna, salmon,
pollock, and catfish.
• Another commonly eaten fish, albacore ("white") tuna has more Hg than canned light tuna. So,
when choosing your two meals offish and shellfish, you may eat up to 6 ounces (one average
meal) of albacore tuna per week.
3. Check local advisories about the safety of fish caught by family and friends in your local lakes, rivers, and
coastal areas. If no advice is available, eat up to 6 ounces (one average meal) per week offish you catch
from local waters, but don't consume any other fish during that week.
Follow these same recommendations when feeding fish and shellfish to your young child, but serve smaller
portions. Source: http://www.fda.gov/bbs/topics/news/2004/NEW01038.html.
http://www.cfsan.fda.gov/~dms/mehgl09.html.
34
-------
Some states with substantial coastlines also issues fish advisories for commercially
marketed fish (e.g., FL, ME), but most appear not to do so. New York State has not issued any
fish advisories for commercially marketed fish, but has issued species-specific freshwater fish
advisories for specific freshwater bodies.
There are several elements to a risk assessment to support development offish advisories:
• Assumed meal size (i.e., 6 oz cooked or 8 oz fresh weight for women in the
EPA/FDA advisory for Hg as MeHg);
• Assumed fish ingestion frequency (e.g., number of meals per week);
• Assumptions concerning which parts offish are eaten (e.g., fillet without skin);
• Monitoring data for chemical concentrations in the fish as marketed (i.e., data
adequacy given spatial and temporal variation and species diversity);
• Assumed body weights of the human receptors (e.g., 65 kg for women);
• Toxicity reference value(s) of concern (e.g., EPA RfD for MeHg); and
• Sensitive lifestages or subpopulations (e.g., pregnant women - developing fetus).
Additional issues are associated with developing fish advisories in narrative form to assist
the public in making informed decisions about purchasing and consuming fish:
• The identification of species of concern by market name and adequacy of
coverage by those names (e.g., the "Spanish Mackerel" in the CM survey is
biologically similar to the "King Mackerel" in the EPA-FDA fish advisory, yet is
not named in the latter fish advisory);
• Providing a conditional narrative that recognizes variation in fish consumption
habits (e.g., how one can select different numbers of servings offish per week
depending on combinations offish in the lower and higher contamination
categories); and
• Accounting for families that also consume self-caught freshwater or estuarine
fish, particularly in regions of the country with relatively high freshwater
contamination levels (e.g., Hg in New England lakes).
To provide a basis of comparison with health-based criteria for Hg concentrations in fish
obtained from the CM, the CM fish tissue concentrations are shown within a risk context using
two types of health-based criteria:
35
-------
1. "Action Levels" expressed as Hg or MeHg concentrations in fish on a wet-weight basis.
This rapidly identifies the fish types that show sufficient Hg residues to merit action by
the agency that published the level depending on their risk management policies.
2. "Allowed Servings per Week" to minimize the possibility that total Hg or MeHg
ingestion by women of child-bearing age would exceed the EPA RfD of 0.1 mg/kg-day
MeHg. This information may assist in developing narrative fish advisories for customers
of the CM.
5.2. COMPARISON OF CM CONCENTRATIONS TO ACTION LEVELS
Table 11 lists several international, U.S. federal, and U.S. state agency standards or
guidelines for total Hg and MeHg residue levels in fish that may trigger an agency action (e.g.,
issuing fish advisories, reducing Hg concentrations in effluents). This table is organized by the
value of the "action level" (in mg [Hg or MeHg]/kg [edible portion fish, wet weight]) and
includes attributes of the level important to its ancillary descriptive information.
Note that the action levels generally are specified as mg [MeHg]/kg [fish edible portion];
in some cases, however, they are specified on the basis of total Hg. The "edible" portion of ray-
finned5 fish generally is the fillet (skinless) (U.S. EPA, 2000a,b), while the edible portion of
other fish may be called other names (e.g., "wings" for skate; "tail" for monkfish, "steak" for
tuna) and the edible portion of shellfish also varies by species (e.g., muscle of scallop; entire
body of clam and mussel).
Fish monitoring programs generally analyze total Hg to minimize costs or to maximize
the number of samples that can be analyzed. EPA recommends that States and Tribes
monitoring fish for Hg levels assume that all Hg in the fish sampled is present as MeHg (U.S.
EPA, 2009). The proportion of total Hg present as MeHg for predatory, or game fish, is high,
more than 90% in most studies (U.S. EPA, 2001, 2009b). The ratio of MeHg to total Hg
generally is less for lower trophic level fish (e.g., trophic-level three fish, approximately 80%;
U.S. EPA, 2009b), and even lower for shellfish (e.g., 49% for oysters from estuaries in South
Carolina, [Kawaguchi et al., 1999]; 35% in blue crabs [Ward et al., 1979]). The EPA considers
it reasonable for States and authorized Native American Tribes to implement the MeHg "Fish
Tissue Residue Criterion" for ambient waters by analyzing fish tissue samples for total Hg first.
They may analyze for MeHg as they may deem necessary. Note that the only governmental
entity for which we identified a different guideline for total Hg than for MeHg is Japan (0.4 and
0.3 mg/kg, respectively). Based on the information in Table 11, we selected four action levels
against which to compare measured total Hg concentrations in fish from the CM, from lowest to
highest: 0.2 mg/kg (Maine), 0.3 mg/kg (EPA), 0.5 mg/kg (Florida/EU/Canada), and 1.0 mg/kg
(FDA).
5 See Table 2, "Class Actinopterygii, Subclass Neopterygii (rayed fish)" for a list of the CM fish in this category.
36
-------
Table 11. International, U.S. federal, and U.S. state MeHg/Hg action levels
Action Level
(mg/kg)a
l.OMeHg
l.OMeHg
0.5 Total Hg
0.5 Total Hg
0.5 Total Hg
0.5 MeHg
0.4 Total Hg
0.3 MeHg
0.3 MeHg
0.2 MeHg
Agency, Office
US Food and Drug
Administration (FDA)b,
CFSANb
FAO/WHO Codex
Alimentarious
Commission
European Union (EU)
Canadian Food
Inspection Agency
Florida Dept. of
Environmental
Protection
FAO/WHO Codex
Alimentarious
Japan
US Environmental
Protection Agency
(EPA), Office of Water
Japan
Maine State Bureau of
Health
Description
Action Level for edible portion
Guideline levels for predatory
fish
Standard for all commercially
sold fish
Standard for all commercially-
sold fish except shark,
swordfish, and fresh/frozen
tuna (next column)
"Safe for unlimited
consumption"
Guideline levels for non-
predatory fish
Guideline for total Hg in fish
Human Health Criteria: MeHg
Fish Tissue Criterion
Guideline for MeHg fish
Fish Tissue Action Level [for
women who are pregnant]
Comments
Compliance policy guides, Sec. 540.600 Fish,
Shellfish, Crustaceans and Other Aquatic
Animals
For shark, swordfish, tuna, pike
Total Hg, with some exceptions
Fish advisory for shark, swordfish, and
fresh/frozen tuna: pregnant women, women of
child-bearing age and young children limit
consumption to no more than one meal/month;
other adults no more than one meal/week
Method/assumptions for derivation of value not
readily available
Guideline for fish other than shark, swordfish,
tuna, pike
Note Japan sets different guidelines for Total Hg
and for MeHg; information based on secondary
citation
Freshwater and estuarine fish, adults of general
population (70 kg), 70% U.S. EPA RfD
apportioned to this source (30% from marine fish
and shellfish)
Note Japan sets different guidelines for Total Hg
and for MeHg; Information based on secondary
citation
Adult females, 60 kg, average long-term fish
ingestion rate 32 g/day
Ref.
(1)
(2)
(2)
(3)
(4)
(2)
(2)
(5,6)
(2)
(7)
"Units of mg[Hg]/kg[fish fresh (wet) weight].
bU.S. FDA = U.S. Food and Drag Administration; CFSAN = Center for Food Safety and Nutrition.
References:
(1) U.S. FDA (2000). Action Levels for Poisonous or Deleterious Substances in Human Food and Animal Feed, Industry Activities Booklet.
August. Accessed 04/14/09 at: http://www.cfsan.fda.gov/~lrd/fdaact.htmlffmerc.
(2) UNEP (United Nations Environment Program) and WHO (World Health Organization) (2008). Guidance for Identifying Populations at Risk
from Mercury Exposure. UNEP DTIE Chemicals Branch and WHO Department of Food Safety, Zoonoses and Foodborne Diseases. Geneva,
Switzerland. August. Accessed on 04/14/09 at: http://www.who.int/foodsafetv/publications/chem/mercurvexposure.pdf.
(3) International Food Law News - FAO/WHO/WTP/Codex - 2006, April 24-28, 2006, 38th Session of the Codex Committee on Food Additives
and Contaminants, The Hague, Netherlands. Accessed on 04/14/09 at: http://www.reading.ac.uk/foodlaw/news/in-06014.htm.
(4) Health Canada (2007). Human Health Risk Assessment of Mercury in Fish and Health Benefits of Fish Consumption. Bureau of Chemical
Safety, Food Directorate, Health Products and Food Branch, Ottawa, Ontario, Canada. Accessed on 04/14/09 at: http://www.hc-sc.gc.ca/fh-
an/pubs/mercur/merc_fish_poisson-eng.php#l.
(5) FDEP (Florida Department of Environmental Protection) (2006). Fish Consumption Health Advisories (last updated October 19, 2006).
Accessed on 04/14/09 at: http://www.dep.state.fl.us/labs/mercurv/docs/fhapre.htm.
(6) U.S. EPA (U.S. Environmental Protection Agency) (2001). Human Health Criteria: Methylmercury Fish Tissue Criterion. Accessed on
04/14/09 at: http://www.epa.gov/waterscience/criteria/methvlmercurv/document.html.
(7) BMH (Maine Bureau of Health) 2001. Maine Bureau of Health Fish Tissue Action Levels. Accessed on 04/14/09 at:
http://www.maine.gov/dhhs/eohp/fish/documents/Action%20Levels%20Writeup.pdf.
37
-------
Tuna (14)
Swordfish (4)
Mahi-mahi (1)
Spanish Mackerel (3)
Halibut (1)
Bluefish (3)
Chilean Sea Bass (1)
Pollock (9)
Monkfish(10)
Porgy (6)
Croaker (9)
Sea Bass (11)
Lobster (1)
Skate (14)
Flounder/Fluke/Sole (15)
Snapper (16)
Catfish (7)
Cod (10)
Whiting (8)
Bass (3)
Mackerel (8)
Ocean Perch (1)
Herring (1)
Oyster (8)
Blue Crab (11)
Tilapia(11)
Squid (12)
Mussel (7)
Rainbow Trout (1)
Clam (7)
Atlantic Salmon (9)
Scallop (7)
Shrimp (7)
I
l i
1
.
'
J
' |
^
j1
=f^
1
1
1
HI
— i
i
i — i
! 1
; — i
! 1
r~i
!_l
a-i
•—
•— •
H
3
3
•
L
I—
H
h
3 —
1
h-
h
h-
h
•
1
_ i
1
i iMpan Concentration
(ND = 1/2 LOD)
™ " Maine Action Level
^^^^U.S. EPA Screening
Level
^^^^Florida Action Level
™ ~ FDA Action Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Mercury Concentration (ppm)
1.0
"Error bars are ± 2 estimated population standard deviations from the mean, if a standard deviation could be determined.
'Numerical value beside market name indicates the sample size (after averaging the duplicates/replicates), including non-detects.
Figure 1. Mean mercury concentrations ± two estimated population
standard deviation for all fish market names and typical action levelsa'b.
38
-------
Figure 1 shows the mean Hg concentrations in each CM market name category compared
to the four action levels, where non-detects have been assumed to equal one-half the reporting
limit. The error bars indicate ± two estimated population S.D. wherever S.D. values could be
calculated. Two S.D. from the mean encompass approximately 95% of the predicted composite
concentrations, assuming the distributions are normal. The numbers in parentheses indicate the
number of composite samples (including non-detects) that were averaged to estimate the mean
Hg concentration for the fish market name category.
The mean Hg concentrations in swordfish and tuna exceeded the Maine action level and
U.S. EPA level of concern, and the Hg concentration in the single mahi-mahi composite sample
was larger than the Maine action level. The Hg concentrations across composite samples for
both tuna and swordfish exhibited wide distributions, indicating that a given composite sample
may or may not have had Hg concentrations above one or more action levels.
To better investigate the relationship between individual composite samples and the
action levels, Figure 3 in Appendix A plots the Hg concentrations in individual composite
samples, ordered from high to low, against the four action levels selected for comparison
(horizontal red lines) for each market name category. The concentrations used in Figure 3 (see
Appendix A) assume that the non-detects are half the reporting limit. All composite samples
which are non-detects are indicated with an "(ND)". None of the total Hg concentrations for a
composite sample for any of the market name categories exceeded the FDA action level for
MeHg. However, 4 of the 14 composite samples (30%) for tuna were larger than the
Florida/EU/Canadian action level of 0.5 mg/kg, and one of those four composite samples was
from bigeye tuna while three were from yellowfm tuna. A total of 10 (70%) of the composite
tuna samples exceeded the EPA screening level and the Maine action level. For swordfish
(market name), two of the four composite samples (50%) were above the Florida/EU/Canadian
action level of 0.5 mg/kg and three (75%) were larger than the EPA guideline and the Maine
action level. The single mahi-mahi composite sample Hg concentration was above the Maine
action level, while one of the three Spanish mackerel composite samples had a measured Hg
concentration equal to the Maine action level.
5.3. ESTIMATES OF THE NUMBER OF SERVINGS PER WEEK OF CM FISH FOR
ADULT FEMALES OF CHILD-BEARING AGE
For each fish market name, we estimated the maximum number of servings (generally
equivalent to "meals") that a woman of child-bearing age/pregnant woman could consume per
week and not exceed a reference toxicity value. A woman of child-bearing age was used to
represent the sensitive population in adults. There are three U.S. federal reference toxicity
values for MeHg, expressed as a daily dose normalized to body weight that each Agency
considers to present minimal risk of adverse developmental effects when mothers are chronically
exposed to MeHg:
39
-------
EPARfD O.lug/kg-day
Agency for Toxic Substances and
Disease Registry (ATSDR)
Minimal Risk Level (MRL) 0.3 ug/kg-day
FDA Action Level 0.5 ug/kg-day
The FDA is in the process of reviewing its action level for MeHg. For the servings-per-
week analysis, the EPA RfD of 0.1 ug [MeHg]/kg-day was used as the reference toxicity value.
The number offish servings per week (SpW) was calculated to be equal to or less than the value
estimated by the following equation:
a „, RfDxBW „ , .
SpW < — x 7 days I week
/"T T^OO
CHg x FSS
where:
SpW = servings per week, often referred to as meal frequency (meals/week),
RfD = EPA reference dose (mg/kg-day) for MeHg,
BW = female body weight (kg),
FSS = fish serving size (kg/serving), and
C#g = concentration of total Hg in edible portion offish (mg/kg wet weight).
For body weight (BW), a range from 65 to 67 kg was used. Various values, generally
between 60 and 70 kg, have been used by federal and state agencies to protect women of child-
bearing age and pregnant women from developmental toxicants. EPA has recommended several
values in the past: 65 kg for women (U.S. EPA, 1993); 65 kg for adult females and 64 kg for
women of reproductive age (U.S. EPA, 1997, 1999); 67 kg (U.S. EPA, 2000c, p. 4-19); and 66
kg for non-lactating and non-pregnant women between the ages of 15 and 44 (i.e., women of
child-bearing age), lactating women, and pregnant women (U.S. EPA, 2004b). EPA's RfD for
MeHg is based on the 67 kg body weight recommended by EPA's Office of Water (U.S. EPA,
2000c) and by EPA's IRIS (www.epa.gov/iris). A value of 65 kg, on the other hand, would yield
slightly more conservative results than using 67 kg in the equation listed above.
For the fish serving size (FSS), we assumed EPA's recommended value of 8 oz (227 g)
fresh weight fish (U.S. EPA, 2000b, 2004a). EPA considers that value equivalent to 6 oz cooked
fish, which is the basis of the joint EPA-FDA advisory illustrated in Section 5.1 (U.S. EPA,
2000a, 2004a). The meal size therefore was set to 0.227 kg/meal.
The concentration of total Hg in fish was set equal to two different values for each market
name category individually: (a) the mean Hg concentration for the category, and (b) the mean
plus two population S.D. of the mean for each category. As discussed above, the concentrations
within two population S.D. of the mean encompass approximately 95% of the estimated
population sample concentrations, assuming the distributions are normal. For this analysis, the
non-detects are assumed to equal half the reporting limit.
40
-------
Table 12 shows the estimated number of servings (meals) per week for an adult female of
child-bearing age that result in intake at or below the EPA level of concern (i.e., the EPA RfD).
The calculations were based on mean and upper 95% confidence limits on mean mercury
concentrations and are listed by species market name in order of decreasing mean Hg
concentration for the market name category. The upper confidence limit is a statistical upper
bound on the actual value of the mean given the data. Use of the upper confidence limit as the
basis for the serving calculation is, in general, a conservative approach since it is greater than the
mean and thus assumes greater consumption. In many cases, however, the number of servings
based on the upper confidence limit is the same as that based on the mean. The calculations
yield estimates that indicate CM tuna, swordfish, and mahi-mahi cannot be eaten on a weekly
basis without incurring an average daily intake above the EPA RfD. (Note that the tuna values
are based on fresh or frozen tuna samples, no canned tuna samples were analyzed.) The estimate
for mahi-mahi, however, is based on a single composite sample and thus is relatively less certain.
Converting to meals per 30-day month, as shown in Table 13, tuna and swordfish could
be eaten twice a month and mahi-mahi could be eaten three times a month and not exceed the
threshold. Using the conservative estimate of Hg concentration (i.e., the mean + two S.D.),
Spanish mackerel, bluefish, Pollock, and monkfish also should not be eaten weekly. Detailed
results of the calculations used to determine the results shown in Table 12 are provided in
Appendix D. These conclusions regarding serving frequency are generally in line with the
current EPA-FDA advisory discussed in Section 5.1, although the latter discusses meal
frequency for canned "light" and canned albacore tuna and does not discuss fresh or frozen tuna
(e.g., tuna steaks, tuna used for sushi).
The analysis indicates that those who eat up to seven meals a week offish or shellfish
should select a diet of low mercury species, such as salmon or scallops. Seven or more servings
per week are under the conservative measure of Hg concentration (mean + two S.D). A more
sophisticated analysis would also incorporate consumption of multiple species in a given week,
although that is beyond the scope of this report. It should be noted that the meal estimates
discussed herein are based on the mercury concentrations measured in a seafood sample obtained
from a commercial market in New York City. These meal estimates are not intended to be
generalizable to the nation's seafood supply.
41
-------
Table 12. Estimated number of fish servings per week for an adult female of
child-bearing age based on means and upper 95% confidence limits on
mercury concentrations by species"
(1)
Market Name of
Species
Tuna
Swordfish
Mahi-Mahi
Spanish Mackerel
Halibut
Bluefish
Chilean Sea Bass
Pollock
Monkfish
Porgy
Croaker
Sea Bass
Lobster
Skate
Flounder
Snapper
Catfish
Cod
Whiting
Bass
Mackerel
(2)
Mean
Mercury
(mg/kg)
0.42*
0.40*
0.22*
0.15
0.15
0.15
0.13
0.13
0.11
0.098
0.084
0.075
0.069
0.060
0.051
0.049
0.044
0.031
0.028
0.025
0.022
(3)
Number of servings per
Week based on Mean
Mercury that result in intake
at or below the EPA level of
concern15
0
0
0
1
1
1
1
1
1
2
2
2
2
3
3
4
4
6
7
8
9
(4)
95% Upper Confidence
Limit on Mean Mercury
(mg/kg)
0.55
0.59
NA
0.20
NA
0.17
NA
0.15
0.13
0.12
0.10
0.088
NA
0.078
0.065
0.060
0.061
0.038
0.043
0.047
0.028
(5)
Number of servings per
Week based on 95% upper
confidence limit on Mean
Mercury that result in
intake at or below the EPA
level of concern15
0
0
NA
1
1
1
NA
1
1
1
2
2
NA
2
3
3
3
5
4
4
7
42
-------
Table 13. Estimated number of fish servings per week for an adult female of
child-bearing age based on means and upper 95% confidence limits on
mercury concentrations by species" (continued)
(1)
Market Name of
Species
Ocean Perch
Herring
Oyster
Blue Crab
Tilapia
Squid
Mussel
Rainbow Trout
Clam
Atlantic Salmon
Scallop
Shrimp
(2)
Mean
Mercury
(mg/kg)
0.022
0.022
0.015
0.015
0.014
0.014
0.012
0.012
0.0081
0.0081
0.0055
0.0054
(3)
Number of servings per
Week based on Mean
Mercury that result in intake
at or below the EPA level of
concern15
9
9
13
13
14
14
16
16
24
24
36
37
(4)
95% Upper Confidence
Limit on Mean Mercury
(mg/kg)
NA
NA
0.025
0.021
0.022
0.017
0.017
NA
0.012
0.012
0.0071
0.0073
(5)
Number of servings per
Week based on 95% upper
confidence limit on Mean
Mercury that result in
intake at or below the EPA
level of concern15
NA
NA
8
9
9
11
11
NA
16
16
28
27
a Serving values calculated using the following exposure assumptions: Serving size = 8 oz offish fresh weight, Adult female
weight = 65 kg, RID for MeHg =1 x 10"4 mg/kg-day (U.S. EPA IRIS database), and the person consumes only the one type of
fish or shellfish.
^Weekly value calculated using the equation for SpWin Section 5.3.
*These concentrations yield values indicating less than one serving a week results in intake at or below the EPA level of concern.
Servings per 30 day month of these species that result in intake at or below the EPA level of concern are provided in Table 13.
NA = Value not available because only one composite sample was analyzed so that a standard deviation required for the
calculation could not be calculated.
43
-------
Table 13. Estimated number of fish servings per 30 day month for an adult
female of child-bearing age based on means and upper 95% confidence limits
on mercury concentrations for three high mercury species
(1)
Market Name
of Species
Tuna
Swordfish
Mahi-Mahi
(2)
Mean
Mercury
(mg/kg)
0.42
0.40
0.22
(3)
Number of servings per 30 Day
MONTH based on Mean
Mercury that result in intake at
or below the EPA level of
concern3
2
2
3
(4)
95% Upper Confidence
Limit on Mean Mercury
(mg/kg)
0.55
0.59
NA
(5)
Number of servings per 30
Day MONTH based on 95%
unner confidence limit on
Mean Mercury that result
in intake at or below the
EPA level of concern"
1
1
NA
aValue per 30 day month calculated using the equation for SpW in Section 5.3, using 30 days rather than 7.
NA = Value not available because only one composite sample was analyzed so that a standard deviation required for the
calculation could not be calculated.
44
-------
6. COMPARISON OF CM DATA TO FDA MONITORING DATA
To further assist in evaluating Hg concentrations measured for composite samples of
seafood from the CM, we provide a comparison of Hg concentrations in fish from the CM with
Hg concentrations for the same market name categories offish as monitored by the FDA from
2000 to 2004 and from 1990 to 2004. This comparison indicates which market names sold in the
CM may be more or less contaminated with Hg than the fish monitored by FDA which are used
to develop national fish advisories for commercially marketed fish.
FDA monitoring data for Hg concentrations in fish are available for download from the
FDA website. For each record in the FDA database, the sample description includes the market
name/common name analyzed, a Hg concentration in mg/kg, and the year of analysis.
Measurements of Hg or MeHg concentrations in fish downloaded from the database were limited
to the years 1995 through 2004. The market names listed in the database were examined and
were then grouped to conform to the CM market names. Table 14 lists the FDA monitoring
"names" that were included under each CM market name, where only market names and FDA
monitoring names that matched are listed. Means and S.D. were calculated for the FDA
monitoring data for two different periods: 1995-2004 to capture as many species as possible and
2000-2004 to focus on the most recent data only.
45
-------
Table 14. Crosswalk between CM market name and FDA monitoring name
CM Market Name
BASS
BLUEFISH
CATFISH
CLAM
COD
CRAB
CROAKER
FLOUNDER/FLUKE/SOLE
HALIBUT
LOBSTER
MACKEREL
MAfflMAffl
OYSTER
PERCH OCEAN
POLLOCK
FDA Monitoring "Sample
Description"
FRESHWATER: BASS
FRESHWATER: BASS
LARGEMOUTH
BLUEFISH
CATFISH
CLAM
COD
COD ALASKAN
COD BLACK
COD GREY
COD PACIFIC
CRAB
CRAB BLUE
CROAKER ATLANTIC
CROAKER WHITE
FLATFISH: FLOUNDER
FLATFISH: PLAICE
ALASKAN
FLATFISH: PLAICE
AMERICAN
FLATFISH: SOLE
FLATFISH: SOLE DOVER
FLATFISH: SOLE PETRALE
FLATFISH: SOLE REX
FLATFISH: SOLE
YELLOWFIN
HALIBUT
LOBSTER
LOBSTER SPINY
MACKEREL
MAfflMAffl
OYSTER
PERCH OCEAN
POLLOCK
CM Market Name
SALMON
SCALLOP
SEA BASS
SEA BASS CHILEAN
SHRIMP
SNAPPER
SWORDFISH
TILAPIA
RAINBOW TROUT
TUNA
WHITING
FDA Monitoring "Sample
Description"
SALMON
SALMON (MIXED SPECIES)
SALMON ATLANTIC
SALMON COHO
SALMON KING (FARMED)
SALMON PINK
SALMON SOCKEYE
SCALLOP
SEA BASS
SEA BASS BLACK
SEA BASS SPOTTED
SEA BASS STRIPED
SEA BASS WHITE
SEA BASS CHILEAN
SHRIMP
SHRIMP PINK
SHRIMP ROCK
SNAPPER
SNAPPER RED
SWORDFISH
TILAPIA
TROUT FRESHWATER
TROUT FRESHWATER (FARMED)
TROUT RAINBOW
(FRESHWATER)
TUNA
TUNA FR/FZN
TUNA FR/FZN ALBACORE
TUNA FR/FZN BIGEYE
TUNA FR/FZN SKIPJACK
TUNA FR/FZN YELLOWFIN
WHITING
46
-------
Table 15, below, provides a comparison between the means and S.D. in the CM
composite samples and the FDA monitoring data by species. The percent differences in the
mean Hg concentrations for CM fish compared with the FDA data as follows:
FDA
so that a negative number indicates the CM mean Hg concentrations are smaller than the FDA
means. The species were broken into four different categories based on trophic level and
taxonomy. The four categories were (1) "Higher Trophic Level Fish"; (2) "Intermediate Trophic
Level Fish"; (3) "Lower Trophic Level Fish, Salmonids, and Squid"; and (4) "Shellfish".
Market names were assigned to one of the four categories using professional judgment and
considering the FDA monitoring concentrations for Hg and basic biological characteristics of the
group. As discussed previously, the CM data were based on composite samples which each
contained multiple individual organisms, and the measured composite Hg concentration
approximated a mean concentration across the individual fish. Thus, comparing the mean across
composite samples was roughly equivalent to comparing the mean concentrations across all fish
that made up the composites.
In general, the CM mean concentrations were lower than the FDA monitoring means
when looking at both the most recent FDA data and the data dating back to 1990. One exception
to this trend was for Pollock, which had much higher concentrations in the nine CM composite
samples than those in the 62 FDA samples. The CM species with the largest Hg
concentrations — tuna — had composite sample mean concentrations that were in fairly good
agreement with the FDA mean concentrations. However, the CM swordfish mean Hg
concentrations were 60-70% smaller than the FDA mean concentrations.
Figure 2 provides a comparison of the mean Hg concentrations between the CM and FDA
measurements in graphical form, where the species were broken into four categories by trophic
levels. The error limits shown are one S.D. above the mean, where the CM population S.D. were
the estimated population standard deviations calculated from the standard deviations across the
composite samples. In nearly all species, the mean CM concentration lied within one S.D. from
the mean FDA concentrations; the only exception were bluefish and swordfish, where the CM
mean concentrations were more than one S.D. lower than the FDA mean concentrations.
47
-------
Table 15. Comparisons of commercial market measured concentrations to FDA monitoring data
a,b
Commercial Fish Market
Num. Comp. Samp.
Mean
Est. Pop. S.D.
FDA, 2000-2004
Num. Samp.
Mean
S.D.
Percent
Difference
FDA, 1995-2004
Num. Samp.
Mean
S.D.
Percent
Difference
Higher Trophic Level Fish
Tuna
Swordfish
Mahi-mahi
Spanish Mackerel
Halibut
Bluefish
Chilean Sea Bass
Monkfish
Croaker
14
4
1
3
1
3
1
10
9
0.42
0.40
0.22
0.15
0.15
0.15
0.13
0.11
0.084
0.39
0.33
N/A
0.078
N/A
0.04
N/A
0.069
0.043
99
17
2
N/A
14
51
39
N/A
35
0.40
1.2
0.43
N/A
0.23
0.34
0.38
N/A
0.072
0.24
0.65
0.035
N/A
0.10
0.13
0.37
N/A
0.036
4.7%
-67%
-48%
N/A
-35%
-57%
-66%
N/A
16%
290
620
22
N/A
46
52
40
N/A
50
0.35
0.98
0.18
N/A
0.25
0.34
0.39
N/A
0.14
0.27
0.51
0.10
N/A
0.23
0.13
0.36
N/A
0.11
19%
-59%
22%
N/A
-40%
-56%
-66%
N/A
-38%
Intermediate Trophic Level Fish
Pollock
Porgy
Sea Bass
Skate
Flounder / Fluke / Sole
Snapper
Catfish
Cod
Bass
Mackerel
9
6
11
14
15
16
7
10
3
8
0.13
0.098
0.075
0.060
0.051
0.049
0.044
0.031
0.025
0.022
0.057
0.04
0.036
0.06
0.051
0.039
0.041
0.019
0.033
0.017
20
N/A
33
N/A
1
37
1
19
N/A
N/A
0.0032
N/A
0.13
N/A
0.015
0.20
0.10
0.084
N/A
N/A
0.0037
N/A
0.08
N/A
N/A
0.29
N/A
0.062
N/A
N/A
4000%
N/A
-43%
N/A
240%
-75%
-56%
-63%
N/A
N/A
51
N/A
46
N/A
23
43
23
39
4
3
0.038
N/A
0.22
N/A
0.045
0.19
0.049
0.095
0.31
0.090
0.12
N/A
0.23
N/A
0.049
0.27
0.084
0.080
0.16
0.085
240%
N/A
-66%
N/A
13%
-74%
-11%
-67%
-92%
-75%
oo
-------
a,b
Table 15. Comparisons of commercial market measured concentrations to FDA monitoring data ' (continued)
Commercial Fish Market FDA,
Num. Comp.
Samp.
Mean
Est. Pop. S.D. Num. Samp.
Lower Trophic Level Fish, Sabnonids, and Squid
Whiting
Ocean Perch
Herring
Tilapia
Squid
Rainbow Trout
Atlantic Salmon
8
1
1
11
12
1
9
0.028
0.022
0.022
0.014
0.014
0.012
0.0081
0.051
N/A
N/A
0.023
0.017
N/A
0.0095
N/A
N/A
N/A
N/A
N/A
30
1
2000-2004
Mean
N/A
N/A
N/A
N/A
N/A
0.077
0.015
c P. Difference
d.U.
FDA, 1990-2004
Num. Samp. Mean S.D.
Difference
N/A
N/A
N/A
N/A
N/A
0.
15
N/A
N/A
N/A
N/A
N/A
N/A
-84%
-46%
2
6
N/A
9
N/A
34
26
0
0.005
N/A
0.01
N/A
0.072
0.026
0
0.012
N/A
0.023
N/A
0.14
0.057
N/A
350%
N/A
45%
N/A
-83%
-69%
Shellfish
Lobster
Oyster
Blue Crab
Mussel
Clam
Scallop
Shrimp
1
8
11
7
7
7
7
0.069
0.015
0.015
0.012
0.0081
0.0055
0.0054
N/A
0.063
0.03
0.048
0.028
0.0088
0.017
10
4
4
N/A
N/A
N/A
N/A
0.22
0.014
0.049
N/A
N/A
N/A
N/A
0.048
0.010
0.012
N/A
N/A
N/A
N/A
-69%
7.9%
-69%
N/A
N/A
N/A
N/A
25
38
51
N/A
6
1
24
0.15
0.013
0.054
N/A
0
0
0.0050
0.099
0.042
0.12
N/A
0
N/A
0.013
-54%
23%
-72%
N/A
N/A
N/A
10%
VO
TDA Monitoring Data downloaded from FDA Website.
bNum. Comp. Samp, is the number of composite samples in the CM data; Num. Samp, is the number of individual samples in the FDA data; Est. Pop. S.D. is the estimated population standard
deviation calculated from the composite samples.
N/A = Value is not available.
-------
Tuna
Ma hi- ma hi
Spanish Mackerel
Halibut
Bluefish
Chilean Sea Bass
Monkfish
Croaker
Higher Trophic Level Fish
1 1 *-i !|
1 ' ^i 1 i
-1=1 1
=1 "
=B— 1
U~i i
1 ' '
1 i
• FDA Monitoring, 1990-2004
™-l • FDA Monitoring, 2000-2004
iPf ' BCM
_LJ 1
0.000 0.500 1.000 1.500 2.000
Mercury Cone (mg/kg)
Intermediate Trophic Level Fish
Pollock
Porgy
Sea Bass
Skate
Flounder / Fluke / Sole
Snapper
Catfish
Cod
Bass
Mackerel
i — i i • hUA Monitoring, lyyu-
r ' ' ' 2004
D FDA Monitoring, 2000-
1 ' ' ' 2004
fe
B — 1
BH
0.000 0.100 0.200 0.300 0.400 0.500 0.600
Mercury Cone (mg/kg)
Figure 2. Bar charts comparing the commercial market mercury
concentrations to the FDA monitoring data concentrations: Bars are sample
means, error limits are one standard deviation above the mean.
50
-------
Lower Trophic Level Fish, Salmonids, and Squid
Whiting
Ocean Perch
Herring
Tilapia
Squid
Rainbow Trout
Atlantic Salmon
• FDA Monitoring, 1990-2004
1 ' D FDA Monitoring, 2000-2004
3 — 1 BCM
™ ,
1
PT1
0.000 0.050 0.100 0.150 0.200 0.250
Mercury Cone (mg/kg)
Shellfish
Lobster
• FDA Monitoring, 1990-
2004
D FDA Monitoring, 2000-
2004
• CM
0.000 0.050 0.100 0.150 0.200 0.250 0.300
Mercury Cone (mg/kg)
Figure 2. Bar charts comparing the commercial market mercury
concentrations to the FDA monitoring data concentrations: Bars are sample
means, error limits are one standard deviation above the mean, (continued)
51
-------
7. QUALITY ASSURANCE/QUALITY CONTROL PROCEDURES
Sample collection, composite preparation, and tissue analyses for this project were
conducted in accordance with the approved project's Quality Assurance Project Plan (QAPP)
entitled, Quality Assurance Project Plan for Sample Collection, Composite, and Analysis for a
Study of Mercury andPCBs in Seafood from the New Fulton Fish Market (May 2008).
Adherence to the QAPP that was developed specifically for this study and to appropriate
laboratory Standard Operating Procedures (SOPs) were evaluated during technical system audits.
The Quality Assurance team consisted of a project QA manager (PQAM) who coordinated with
each QA manager responsible for oversight of each project task. A quality assurance audit of
each laboratory portion of the project was conducted by the QA manager associated with each
laboratory. Data validation was accomplished through comparison of paper records from field
record forms, chain-of-custody records, processing forms, and logs with output from computer
files and internal QC checks on the study's database. The project's data manager maintained
oversight of the data collected and analyzed and linked the laboratory results in a single project
database.
All laboratory QC procedures set forth in Section 14 of both SOP's (C-104 for PCB's,
and C-l 10 for Mercury) were followed. The analytical laboratory that conducted the mercury
and PCB analyses used the LEVIS system and exported the results to the project data manager for
loading into the project database. For quality control of the DNA analyses, please refer to
Appendix B.
A sample from a small piece of tissue from each of the individual specimens was shipped
for DNA typing to verify the fish species. One sample tissue per composite was DNA typed and
the others were archived. For species that were obtained from the FFM in fillet form, DNA
typing was performed on all three fillets that comprised the composite. In addition, a tissue
archive sample from each individual specimen as well as a composite archive sample from each
of the composite samples was created and archived. The samples that were used for duplicate
analysis were selected according to a systematic random sampling procedure.
Sample analyses conducted for this report were performed in accordance with the Quality
Assurance Project Plan for Data Analysis of Mercury andPCBs in Seafood from the New Fulton
Fish Market (2009). This included a check of the underlying raw data, validation of data
transfer, an assessment of measurement error, and verification of the calculations.
Section 4 of this report describes the assessment of measurement error conducted using the
results of the duplicate samples. The measurement of error variance was calculated for each of
the samples using the formula in Section 4.
7.1. QA/QC OF THE DATA DELIVERY
The following steps were taken to ensure accurate delivery and understanding of the database
contents:
52
-------
1. Ensuring that all data fields listed in the Data Description sheet existed in the
appropriate table.
2. Ensuring that each field contained values that were all of the correct field type (text,
double precision numbers, integers, or dates).
3. Ensuring that fields that could take on only certain discrete values contained only those
values (e.g., "Detected" must be "Y" or "N", "QC_Type" must be "Screen" or "Quant").
4. Verifying that all laboratory results reported were for the analyte Mercury
5. Ensuring that all sample groups were represented on each of the three data tabs.
6. Checking that all text fields contained entries that could be interpreted by the analysts.
7. Checking that all numerical fields contained values that lied within expected ranges and
with consistent significant figures, given the reported units and reported species.
8. Checking that all reported Hg levels indicated as "detected" had values above the
reported detection limit.
Following the database review, the data were considered usable for further data analysis
with the following issues noted and resolved:
• In the "Reporting_Limit" column in Lab Mercury Results, approximately one-third of
the values were equal to the method reporting limit listed in the sample collection and
analysis QAPP (0.02 mg/kg), while the other two-thirds were approximately half this
value. It was confirmed that these were sample-specific reporting limits and were
determined using the sample weight and the lowest calibration point on the analytical
calibration curve. For the analysis, sample-specific reporting limits were used to estimate
non-detected Hg concentrations.
• Upon investigating the "Reporting_Limit" in the Lab Mercury Results sheet as part of
step 7, one value was flagged. The entry on row 217 (Sample 028.C, Lab Sample Name
AK04148) listed a reporting limit of exactly zero, while all other reporting limits were
above zero as expected. It was confirmed that the value was above zero but was not
entered in the database because a detectable amount of Hg was found in the sample.
Because analysis of this particular sample yielded a detectable Hg level, the reporting
limit was not needed for the analysis.
• Upon investigating the "Result_Qualifier" and "Lab_Result_Qualifier" columns in
Lab Mercury Results as part of step 6, all entries were observed to be blank. No entries
could indicate:
Case 1. Data were censored prior to transmission into the Excel spreadsheet.
Case 2. None of the data were transferred for that particular data field—in which
case these fields would need to be updated.
Case 3. All of the cells turned out to be blank for Hg.
53
-------
• It was confirmed that a qualifier was not needed since only values that were above the
reporting limit were entered into the database. Based on this assumption, Case 3 applies.
• In the Lab Mercury Results tab, all entries in the "Basis" column say "N/A". The
possible entries in this column are "wet", "dry", and "N/A". The results were actually
wet weight basis.
• In the Field Lab Processing sheet, it was unclear what the entries in the "WholePart"
column indicate (e.g., "Whole" and "Partial"). It was confirmed that "Whole" meant the
whole entire fish, or most of the entire fish, was collected as a sample specimen at the
market. "Partial" indicated only a part or portion of the fish, such as a fillet or part of a
loin, was collected at the market, prior to homogenization.
• In the Field Collection sheet, the entries in the "Form" column indicated the size of the
original sample (e.g., "Whole gutted", "Whole", and "Fillet"). It was confirmed that an
entry could say "Whole" even if it has been gutted.
To check the species identification as part of step 6, autofilters were applied to all
columns, then the rows in the Lab Mercury Results sheet were sorted by family, genus, and
species. Spelling of family, genus, and species names was checked, and the data were sorted
again after correcting spelling. The following corrections were made:
Lonigo > Loligo for Atlantic squid genus,
pealei >pealeii for Atlantic squid species,
Pluronectidae > Pleuronectidae for several flatfish records,
Merlucciiidae > Merlucciidae for whiting family,
Sebastidae > Serranidae for ocean perch family,
Sebastes > Serranus for ocean perch genus, and
marinus > scriba for ocean perch species.
It was noted thatMorone chrysopes x saxatilis is listed as an "invalid" species name in
the Integrated Taxonomy Information System (ITIS) used by EPA and other federal agencies.
The currently valid genus and species name in ITIS for the ocean perch is Serranus scriba, not
Sebastes marinus, although this was not necessary for analysis.
7.2. QA/QC OF THE ANALYSIS
Commercially available software was used for data analysis. The algorithms used in the
calculations were verified for accuracy prior to their use for sample data analysis. Calculation
reviews focused on correct transcription of equations and correct input ranges/data sources in
addition to evaluation of the statistical approach, compliance with the QAPP, technical validity,
and reasonableness of the results.
54
-------
A two-pronged approach was used to check the calculations performed to estimate the statistics
and risk-based estimates for this report. First, each calculation was performed independently by
two different analysts. The analysts worked from a common template but did not discuss the
analysis prior to beginning the analysis. They used independent analytical tools, including a
combination of Microsoft Access®, Microsoft Excel®, and Visual Basic for Applications®.
Each calculation was then compared to ensure the calculations agreed to within 0.01%. Next, the
QA Officer examined all the calculations, checked for adherence to significant figures, and
determined satisfactory completion of the analysis. The QA Officer also examined the report to
check for accuracy and internal consistency.
55
-------
8. REFERENCES
Burger, J; Stern, AH; Dixon, C; Jeitner, C; Shukla, S; Burke, S; Gochfeld, M (2004). Fish
availability in supermarkets and fish markets in New Jersey. Sci Total Environ 333(1-
3):89-97.
Fabrizio, MC; Frank, AM; Savino, JF (1995). Procedures for Formation of Composite Samples
from Segmented Populations. Environ Sci Technol 29(5): 1137-1144.
Gilbert, RO (1987). Statistical Methods for Environmental Pollution Monitoring, New York,
NY: Wiley and Sons.
Kawaguchi, T; Porter, D; Bushek, D; Jones, B (1999). Mercury in the American oyster
Crassostrea virginica, in high salinity salt marsh estuaries in South Carolina, USA and
the public health concern for the future. Mar Pollut Bull 38:324-327.
McKelvey, W; Gwynn, RC; Jeffery, N; Kass, D; Thorpe, LE; Garg, RK; Palmer, CD; Parsons,
PJ (2007). A biomonitoring study of lead, cadmium, and mercury in the blood of New
York City adults. Environ Health Perspect 115(10):1435-1441.
McKelvey, W; Chang, M; Arnason, J; Jeffrey, N; Kricheff, J; Kass, D (2010). Mercury and
polychlorinated biphenyls in Asian market fish: A response to results from mercury
biomonitoring in New York City. Environ Res 110(7):650ket.
U.S. EPA (U.S. Environmental Protection Agency) (1993). Guidance for Assessing Chemical
Contaminant Data for Use in Fish Advisories. Volume 1, Fish Sampling and Analysis,
Second Edition. Office of Water, Washington, D.C. EPA 823-R-93-002.
U.S. EPA (U.S. Environmental Protection Agency) (1997). Guidance for Assessing Chemical
Contaminant Data for Use in Fish Advisories. Volume 2, Risk Assessment and Fish
Consumption Limits, Second Edition. Office of Water, Washington, D.C. EPA 823-B-97-
009.
U.S. EPA (U.S. Environmental Protection Agency) (1999). Guidance for Assessing Chemical
Contaminant Data for Use in Fish Advisories. Volume 1, Fish Sampling and Analysis,
Third Edition, DRAFT. Office of Water, Washington, D.C. EPA 823-R-99-007.
U.S. EPA (U.S. Environmental Protection Agency) (2000a). Guidance for Assessing Chemical
Contaminant Data for Use in Fish Advisories. Volume 1, Fish Sampling and Analysis,
Third Edition. Office of Water, Washington, D.C. EPA 823-B-00-007. Available at:
http://www.epa.gov/waterscience/fish/advice/volumel/.
U.S. EPA (U.S. Environmental Protection Agency) (2000b). Guidance for Assessing Chemical
Contaminant Data for Use in Fish Advisories. Volume 2, Risk Assessment and Fish
Consumption Limits, Third Edition. Office of Water, Washington, D.C. EPA 823-B-OO-
008. Available at: http://www.epa.gov/waterscience/fish/advice/volume2/.
56
-------
U.S. EPA (U.S. Environmental Protection Agency) (2000c). Methodology for Deriving Ambient
Water Quality Criteria for the Protection of Human Health Final. Office of Water,
Washington, D.C. EPA 822-B-00-004. Available at:
http://www.epa.gov/waterscience/criteria/humanhealth/method/index.html.
U.S. EPA (U.S. Environmental Protection Agency) (2001). Water Quality Criterion for the
Protection of Human Health: Methylmercury. Office of Water, Washington, D.C. EPA
823-R-01-001. Available at:
http://www.epa.gov/waterscience/criteria/methvlmercurv/pdf/mercury-criterion.pdf.
U.S. EPA (U.S. Environmental Protection Agency) (2004a). Technical Memorandum: Origin of
1 Meal/Week Noncommercial Fish Consumption Rate in National Advisory for Mercury.
March 11, 2004. Office of Water, Washington, D.C. EPA 823-B-00-007. Available at:
http://www.epa.gov/waterscience/fish/advice/l-meal-per-week.pdf.
U.S. EPA (U.S. Environmental Protection Agency) (2004b). Estimated Per Capita Water
Ingestion and Body Weight in the United States - An Update. Office of Water, Office of
Science and Technology, Washington, D.C. EPA-822-R-00-001. Available at:
http://www.epa.gov/waterscience/criteria/drinking/percapita/2004.pdf.
U.S. EPA (U.S. Environmental Protection Agency) (2008). Quality Assurance Project Plan for
Sample Collection, Composite, and Analysis for a Study of Mercury and PCBs in
Seafood from the New Fulton Fish Market. Office of Research and Development,
Washington, D.C.
U.S. EPA (U.S. Environmental Protection Agency) (2009a). Quality Assurance Project Plan for
Data Analysis of Mercury and PCBs in Seafood from the New Fulton Fish Market. Office
of Research and Development, Washington, D.C.
U.S. EPA (U.S. Environmental Protection Agency) (2009b). Guidance for Implementing the
January 2001 Methylmercury Water Quality Criterion. Office of Water, Office of Science
and Technology. Washington, D.C. EPA 823-R-09-002. Accessed February 14:
http: //www. epa.gov/waterscience/criteria/methylmercury/guidance -final .html.
Ward, DR; Nickelson II, R; Finne, G (1979). Relationship between methylmercury and total
mercury in blue crabs (Callinectes sapidus). J Food Sci 44(3):920-921.
57
-------
APPENDIX A. DETAILED STATISTICAL ANALYSIS RESULTS
Table A-l. Statistical information by market name, non-detects equal to the reporting limit
a,b
Market
Name of
Species
Tuna
Swordfish
Mahi-mahi
Mackerel,
Spanish
Halibut
Bluefish
Bass, Chilean
Sea
Pollock
Monkfish
Porgy
Croaker
Bass, Sea
Lobster
Skate
Flounder/
Fluke/Sole
Snapper
Number of
Composite
Samples0
Det.
14
4
1
3
1
3
1
9
10
6
9
11
1
13
13
16
N.D.
0
0
0
0
0
0
0
0
0
0
0
0
0
1
2
0
Tot.
14
4
1
3
1
3
1
9
10
6
9
11
1
14
15
16
Perc.
N.D.
(%)
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
7%
13%
0%
Mean
(mg/kg)
0.42
0.40
0.22
0.15
0.15
0.15
0.13
0.13
0.11
0.098
0.084
0.075
0.069
0.060
0.051
0.049
Comp.
S.D.
(mg/kg)
0.25
0.19
N/A
0.045
N/A
0.023
N/A
0.034
0.044
0.023
0.024
0.021
N/A
0.034
0.027
0.022
Est. Pop.
S.D.
(mg/kg)
0.39
0.33
N/A
0.078
N/A
0.04
N/A
0.057
0.069
0.040
0.043
0.036
N/A
0.059
0.049
0.039
Est.
Pop.
C.V.
(%)
93%
82%
N/A
51%
N/A
27%
N/A
44%
65%
41%
51%
49%
N/A
97%
96%
80%
Min.
(mg/kg)
0.043
0.14
N/A
0.11
N/A
0.12
N/A
0.079
0.054
0.068
0.056
0.03
N/A
0.0099
0.0093
0.017
25th
Perc.
(mg/kg)
0.23
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.073
N/A
N/A
0.064
N/A
0.03
0.038
0.032
Median
(mg/kg)
0.39
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.095
N/A
N/A
0.078
N/A
0.064
0.049
0.044
75th Perc.
(mg/kg)
0.57
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.14
N/A
N/A
0.087
N/A
0.081
0.066
0.068
Max.
(mg/kg)
0.82
0.57
N/A
0.2
N/A
0.16
N/A
0.18
0.18
0.13
0.13
0.11
N/A
0.12
0.1
0.083
Est. Lower
95% C.L.
on Meand
0.29
0.22
N/A
0.1
N/A
0.12
N/A
0.11
0.08
0.079
0.069
0.063
N/A
0.042
0.038
0.038
Est.
Upper
95%
C.L. on
MeanD
0.55
0.59
N/A
0.2
N/A
0.17
N/A
0.15
0.13
0.12
0.1
0.088
N/A
0.078
0.065
0.06
-------
Table A-l. Statistical information by market name, non-detects equal to the reporting limit' (continued)
a,b
Market
Name of
Species
Catfish
Cod
Whiting
Bass
Mackerel
Perch, Ocean
Herring
Tilapia
Oyster
Crab, Blue
Squid
Mussel
Trout,
Rainbow
Salmon,
Atlantic
Clam
Scallop
Shrimp
Number of
Composite
Samples0
Det.
7
10
7
3
7
1
1
5
6
8
10
5
1
3
3
1
1
N.D.
0
0
1
0
1
0
0
6
2
3
2
2
0
6
4
6
6
Tot.
7
10
8
o
6
8
1
1
11
8
11
12
7
1
9
7
7
7
Perc.
N.D.
(%)
0%
0%
13%
0%
13%
0%
0%
55%
25%
27%
17%
29%
0%
67%
57%
86%
86%
Mean
(mg/kg)
0.044
0.031
0.029
0.025
0.023
0.022
0.022
0.017
0.017
0.016
0.015
0.014
0.012
0.011
0.011
0.0094
0.0093
Comp.
S.D.
(mg/kg)
0.023
0.012
0.020
0.019
0.0077
N/A
N/A
0.011
0.013
0.0077
0.0051
0.0037
N/A
0.0035
0.0042
0.00074
0.0009
Est. Pop.
S.D.
(mg/kg)
0.041
0.019
0.049
0.033
0.017
N/A
N/A
0.02
0.059
0.025
0.014
0.031
N/A
0.006
0.02
0.0029
0.0061
Est.
Pop.
C.V.
(%)
93%
63%
170%
130%
74%
N/A
N/A
110%
350%
150%
93%
230%
N/A
54%
190%
31%
66%
Min.
(mg/kg)
0.024
0.016
0.0096
0.014
0.013
N/A
N/A
0.0092
0.0092
0.0085
0.0084
0.0087
N/A
0.0086
0.0085
0.0086
0.0084
25th
Perc.
(mg/kg)
N/A
0.024
N/A
N/A
N/A
N/A
N/A
0.0097
N/A
0.009
0.011
N/A
N/A
N/A
N/A
N/A
N/A
Median
(mg/kg)
N/A
0.027
N/A
N/A
N/A
N/A
N/A
0.0099
N/A
0.017
0.014
N/A
N/A
N/A
N/A
N/A
N/A
75th Perc.
(mg/kg)
N/A
0.038
N/A
N/A
N/A
N/A
N/A
0.021
N/A
0.023
0.017
N/A
N/A
N/A
N/A
N/A
N/A
Max.
(mg/kg)
0.094
0.049
0.075
0.047
0.034
N/A
N/A
0.038
0.047
0.029
0.024
0.019
N/A
0.019
0.02
0.011
0.011
Est. Lower
95% C.L.
on Meand
0.026
0.024
0.014
0.0034
0.017
N/A
N/A
0.011
0.0078
0.012
0.012
0.011
N/A
0.0089
0.0075
0.0089
0.0086
Est.
Upper
95%
C.L. on
MeanD
0.061
0.038
0.043
0.047
0.028
N/A
N/A
0.023
0.025
0.021
0.017
0.016
N/A
0.013
0.014
0.0099
0.01
>
-------
Table A-l. Statistical information by market name, non-detects equal to the reporting limita'b (continued)
Market
Name of
Species
Total
Number of
Composite
Samples0
Det.
194
N.D.
42
Tot.
236
Perc.
N.D.
(%)
18%
Mean
(mg/kg)
0.076
Comp.
S.D.
(mg/kg)
0.12
Est. Pop.
S.D.
(mg/kg)
N/A
Est.
Pop.
C.V.
(%)
N/A
Min.
(mg/kg)
0.0084
25th
Perc.
(mg/kg)
0.014
Median
(mg/kg)
0.034
75th Perc.
(mg/kg)
0.083
Max.
(mg/kg)
0.82
Est. Lower
95% C.L.
on Meand
0.06
Est.
Upper
95%
C.L. on
MeanD
0.092
aDet. is the number of samples with results above the reporting limit (referred to here as "detects"), N.D. is the number of samples with results below the reporting limit,
Tot. is the total number of samples for the species (excluding replicate and duplicate samples), Perc. N.D. is the percent of the total number of samples which are below the
reporting limit, Comp. S.D. is the standard deviation of the mean of the composite sample Hg concentrations for the market name, Est. Pop. S.D. is the estimated population
standard deviation, Est. Pop. C.V. is the estimated coefficient of variation calculated as the population standard deviation/mean x 100%, 25th Perc. is the 25th percentile, 75th
Perc is the 75th percentile, Est. Lower 95% C.L. on mean is the estimated lower 95th percent confidence limit on the mean, and Est. Upper 95% C.L. on Mean is the
estimated upper 95th percent confidence limit on the mean.
bNumber of Market Name groups included =33.
°Number of composite samples after averaging the duplicates and replicates for a single composite sample and treating the average as a single point.
dCalculated based on the estimated population standard deviation.
N/A = Value not available.
>
-------
Table A-2. Mercury concentrations by species or by condition (e.g., hard- or softshell) within a market name; non-
detects equal to the reporting limita'b
Market Name
Catfish
Codd
Blue Crab
Flounder/ Fluke/
Sole
Shrimp
Species
Catfish, Blue
Catfish, Channel
Catfish, White
All Catfish
Cod, Atlantic
Cod, Pacific
All Cod
Blue Crab/Hardshell
Blue Crab/Softshell
All Crab
Flounder, Blackback
Flounder, Summer
Sole, Gray
All Flounder/Fluke/Sole
Shrimp, Black Tiger
Shrimp, White
All Shrimp
Number of Composite
Samples0
Del.
2
4
1
7
7
3
10
5
3
8
4
5
4
13
0
1
1
N.D.
0
0
0
0
0
0
0
0
3
3
1
0
0
2
2
4
6
Total
2
4
1
7
7
3
10
5
6
11
5
5
4
15
2
5
7
Perc. N.D.
(%)
0%
0%
0%
0%
0%
0%
0%
0%
50%
27%
20%
0%
0%
13%
100%
80%
86%
Mean
(mg/kg)
0.037
0.048
0.041
0.044
0.031
0.030
0.031
0.021
0.012
0.016
0.032
0.073
0.059
0.051
0.009
0.010
0.009
Comp. S.D.
(mg/kg)
0.012
0.032
N/A
0.023
0.013
0.009
0.012
0.003
0.008
0.008
0.015
0.026
0.012
0.027
0.000
0.001
0.001
Est. Pop. S.D.
(mg/kg)
0.021
0.055
N/A
0.041
0.022
0.015
0.019
0.013
0.024
0.025
0.028
0.045
0.020
0.049
0.001
0.006
0.006
Est. Pop. C.V.
(%)
57%
110%
N/A
93%
70%
49%
63%
62%
190%
150%
88%
61%
34%
96%
13%
64%
66%
Min.
(mg/kg)
0.028
0.024
N/A
0.024
0.016
0.024
0.016
0.017
0.009
0.009
0.010
0.045
0.044
0.009
0.008
0.009
0.008
Max.
(mg/kg)
0.045
0.094
N/A
0.094
0.049
0.040
0.049
0.025
0.029
0.029
0.049
0.100
0.071
0.100
0.009
0.011
0.011
>
-------
Table A-2. Mercury concentrations by species or by condition (e.g., hard- or softshell) within a market name; non-
detects equal to the reporting limita'b (continued)
Market Name
Snapper
Squidd
Tuna
Whiting
Species
Snapper, Caribbean Red
Snapper, Lane
Snapper, Red
Snapper, Vermilion
Snapper, Yellowtail
All Snapper
Squid, Japanese Flying
Squid, Longfin (Atlantic)
All Squid
Yellowfin Tuna
Bigeye Tuna
All Tuna
Whiting, Offshore
Whiting/Silver Hake
All Whiting
Number of Composite
Samples0
Del.
1
2
6
O
4
16
3
7
10
7
7
14
2
5
7
N.D.
0
0
0
0
0
0
1
0
2
0
0
0
0
1
1
Total
1
2
6
O
4
16
4
7
12
7
7
14
2
6
8
Perc.
N.D.
(%)
0%
0%
0%
0%
0%
0%
25%
0%
17%
0%
0%
0%
0%
17%
13%
Mean
(mg/kg)
0.039
0.040
0.057
0.040
0.051
0.049
0.012
0.017
0.015
0.420
0.410
0.420
0.052
0.021
0.029
Comp. S.D.
(mg/kg)
N/A
0.018
0.020
0.037
0.022
0.022
0.003
0.005
0.005
0.280
0.230
0.250
0.032
0.009
0.020
Est. Pop. S.D.
(mg/kg)
N/A
0.031
0.035
0.065
0.038
0.039
0.006
0.017
0.014
0.440
0.360
0.390
0.062
0.025
0.049
Est. Pop. C.V.
(%)
N/A
78%
63%
160%
76%
80%
47%
100%
93%
100%
87%
93%
120%
120%
170%
Min.
(mg/kg)
N/A
0.027
0.031
0.017
0.034
0.017
0.009
0.012
0.008
0.043
0.130
0.043
0.029
0.010
0.010
Max.
(mg/kg)
N/A
0.052
0.076
0.083
0.082
0.083
0.015
0.024
0.024
0.800
0.820
0.820
0.075
0.033
0.075
>
"Column names are as defined in Table A-l.
'Number of Market Names encompassing multiple species or conditions equals 10 of the 33 total Market Names.
'Number of composite samples after averaging the duplicates and replicates for a single composite sample and treating the average as a single point.
''Species from different oceans.
N/A = Value not available because only one composite was analyzed.
-------
Table A-3. Mercury concentrations by water body of origin within a market name species; non-detects equal to the
reporting limit"
Market
Name of
Species
Clam
Cod
Mussel
Salmon
Snapper
Squid
Water Body or Water Type
of Origin
Clam, Atlantic, Wild
Clam, Farmed
Clam, Sound
All Clam
Cod, Atlantic, Wild
Cod, Pacific, Wild
All Cod
Mussel, Atlantic, Wild
Mussel, Farmed
All Mussel
Atlantic Salmon, Wild
Atlantic Salmon, Farmed
All Salmon
Snapper, Atlantic, Wild
Snapper, Pacific, Wild
Snapper, Unknown
All Snapper
Squid, Atlantic, Wild
Squid, Pacific, Wild
All Squid
Number of Composite
Samples"
Del.
1
2
0
3
7
3
10
4
1
5
0
3
3
13
2
1
16
7
3
10
N.D.
3
0
1
4
0
0
0
1
1
2
3
3
6
0
0
0
0
0
2
2
Total
4
2
1
7
7
3
10
5
2
7
3
6
9
13
2
1
16
7
5
12
Perc.
N.D.
(%)
75%
0%
100%
57%
0%
0%
0%
20%
50%
29%
100%
50%
67%
0%
0%
0%
0%
0%
40%
17%
Mean
(mg/kg)
0.011
0.0095
0.0094
0.011
0.031
0.03
0.031
0.014
0.012
0.014
0.0089
0.012
0.011
0.05
0.04
0.047
0.049
0.017
0.011
0.015
Comp. S.D.
(mg/kg)
0.0057
0.00071
N/A
0.0042
0.013
0.0086
0.012
0.0037
0.0045
0.0037
0.00043
0.0038
0.0035
0.024
0.018
N/A
0.022
0.0048
0.0025
0.0051
Est. Pop.
S.D.
(mg/kg)
0.027
0.0041
N/A
0.02
0.022
0.015
0.019
0.031
0.038
0.031
0.00075
0.0066
0.006
0.042
0.031
N/A
0.039
0.017
0.0054
0.014
Est. Pop.
C.V. (%)
230.0%
43.0%
N/A
190.0%
70.0%
49.0%
63.0%
220.0%
320.0%
230.0%
8.5%
53.0%
54.0%
83.0%
78.0%
N/A
80.0%
100.0%
49.0%
93.0%
Min.
(mg/kg)
0.0085
0.009
N/A
0.0085
0.016
0.024
0.016
0.0088
0.0087
0.0087
0.0086
0.0088
0.0086
0.017
0.027
N/A
0.017
0.012
0.0084
0.0084
Max. (mg/kg)
0.02
0.01
N/A
0.02
0.049
0.04
0.049
0.019
0.015
0.019
0.0094
0.019
0.019
0.083
0.052
N/A
0.083
0.024
0.015
0.024
-------
Table A-3. Mercury concentrations by water body of origin within a market name species; non-detects equal to
the reporting limit" (continued)
Market
Name of
Species
Swordfish
Water Body or Water Type
of Origin
Swordfish, Atlantic, Wild
Swordfish, Pacific, Wild
All Swordfish
Number of Composite
Samples"
Del.
2
2
4
N.D.
0
0
0
Total
2
2
4
Perc.
N.D.
(%)
0%
0%
0%
Mean
(mg/kg)
0.33
0.48
0.4
Comp. S.D.
(mg/kg)
0.26
0.13
0.19
Est. Pop.
S.D.
(mg/kg)
0.45
0.22
0.33
Est. Pop.
C.V. (%)
140.0%
46.0%
82.0%
Min.
(mg/kg)
0.14
0.39
0.14
Max. (mg/kg)
0.51
0.57
0.57
"Column names are as defined in Table A-l.
bNumber of composite samples after averaging the duplicates and replicates for a single composite sample and treating the average as a single point.
N/A = Value not available because only one composite was analyzed.
>
-------
132
Q
% 195
o
is
098
0.
1
4
4
4
^
4
4
D 0.1 0
> i
> i
> i
> i
> i
i
> i
> i
2 0
I/
i
i
i
i
i
i
i
i
It
Bass
i
, i
, i
i
i
i
i
i
i
i
i
i
i
i
i
i
^^•Average Composite
Sample Cone.
-^"Maine Action Level
—^— US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
— ^US FDA Action
Level
3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Mercury Concentration (ppm)
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels.
A-8
-------
Chilean Sea Bass
Q.
3
o
3
069
^
4
<
^
4
<
> 1
> 1
> 1
> 1
> 1
> 1
> 1
*, r
1 A
\ A
\ A
\ A
\ A
\ A
\ A
1 , /
<
k 1
<
1
1 1
1 1
1
1
1
1
1
1
^•Average Composite
Sample Cone.
•^"Maine Action Level
—^— US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
•^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8
0.9
1.0
Sea Bass
075
096
025
-
120
-
099
= 250
o
O
054
-
155
-
129
164
012
1
HI
3
— 1 4
4
4
4
4
4
> I
> I
> I
> I
> |
> I
> I
> I
•> r
I A
I A
I A
I A
| i
I A
I A
I A
1 , t
(
(
<
(
^
(
«
<
r
i
i
1
i
i
i
|fl
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8
0.9
1.0
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-9
-------
Bluefish
154
0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8 0.9
• Average Composite
Sample Cone.
•Maine Action Level
•US EPA Screening
Level
•Florida/EU/Canada
Action Level
•US FDA Action
Level
1.0
166
101
Catfish
0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8 0.9
• Average Composite
Sample Cone.
•Maine Action Level
•US EPA Screening
Level
•Florida/EU/Canada
Action Level
•US FDA Action
Level
1.0
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-10
-------
Clam
040
199
064
Q
= 033
0
O
173
193
043
] <
4
4
4
(ND)
4
(ND)
(ND) <
4
(ND)
f \
> I
> I
> I
> I
> I
> I
> I
s r
j 1
1 i
\ i
\ i
\ *
\ i
It
i
\ i
\ i
~\ , t
l v
<
(
(
<
(
|
L V
<
<
f
1
1
1
1
|
|
1
i lAvpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
-^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
Cod
0.8
0.9
1.0
077 1
062 1
222 H <
050 l| <
tr
Q 187 F <
= -c
o 1— |
0 224 | <
004 11 .
223 fj
163 1
I — 4
252 I
l : f
f \
> 1
> 1
> 1
> 1
> 1
> 1
> |
s r
j 1
1 i
\ i
\ i
\ *
\ i
II
i
\ i
\ i
~\ , t
l v
<
(
(
<
(
\ t
L V
<
{
f
1
1
1
1
|
|
1
i lAvpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8
0.9
1.0
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-ll
-------
Blue Crab
065
169
260
081
133
9
§" 145
o
5
244
242
207
144
171
1
1
4
4
4
ND) <
(ND)
4
(ND)
r i
> \
> \
> \
> \
> \
> \
> \
> \
v r
j j
1 A
\ A
\ A
\ A
\ A
\ A
\ A
\ A
1
i V
i
i i
(
L i
i i
i
i i
i i
f
1
1
1
1
1
1
1
1
^•Average Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Mercury Concentration (ppm)
116
023
128
^
4
153 ^|
0
= 215
0
0
209
-
1
4
^
^
1
105 I 1
213 1
I ' >
i <
091 I
> \
> \
> \
> \
> \
> \
> \
v r
J A
\ A
\ A
\ A
\ A
\ A
\ A
\ A
A
\
Croaker
\ r
t 1
k 1
1
k 1
t 1
1
k (
k <
1
1
1
1
1
1
1
1
I lAvpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Mercury Concentration (ppm)
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-12
-------
Flounder, Fluke, and Sole
112 j |
061 j | <
103 •""!
122 | <
031 j |
076 ^J
047 ^] <
138 t]
046 ^f]
048 P
001 1(ND) <
002 1 (ND)
r 1
> 1
> 1
> |
> 1
> 1
> 1
> 1
v r
J j
1 i
\ 1
\ <
1 .
1
\ i
\ i
\ i
-\
; ^
(
<
L c
<
<
1
<
<
r
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
i
i
i
i
i
i
i
0
I lAwpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
Mercury Concentration (ppm)
Halibut
g
= 074
o
o
*
4
4
4
4
4
4
4
r 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
s r
j j
1 i
\ *
1 ,
\ .
1 i
\ i
\ i
\ i
i , t
-t ^
(
<
i
i
I
(
(
<
(
(
r
i
i
i
i
i
i lAwpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8
0.9
1.0
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-13
-------
Herring
g
= 088
o
o
4
4
4
4
4
4
4
4
/ 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
s r
j ?
1 i
\ *
1 i
\ i
\ i
\ i
\ i
\ i
i , t
: ^
(
<
i
i
I
(
(
<
(
(
r
i
i
i
i
i
i lAwpragp Composite
Sample Cone.
-^"Maine Action Level
—^— US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
Lobster
0.8
0.9
1.0
g
= 146
o
O
4
4
4
4
4
4
4
4
r 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
s r
j j
1 i
\ *
1 ,
\ i
\ i
\ i
\ i
\ i
i , t
-t ^
(
<
i
i
I
(
(
<
(
(
r
i
i
i
i
i
i lAwpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Mercury Concentration (ppm)
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-14
-------
Mackerel
210 1
1 — ^
212 1
211 I
-T
Q 235 1 <
0. J— '
0 239 1 <
_M
Vi
240 1
237 Q(ND)
020 1
f 1
> 1
> 1
y r
f i
> 1
> 1
> 1
> 1
v r
j j
I i
\ A
II
A
\ A
\ A
Ik
1
\ i
\ A
I
k V
<
, 1
V i
L V-
L 1
, (
L (
, 1
, 1
("
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
\
I
|
1
I
I
1
0
^^Average Composite
Sample Cone.
CZNon-Detect (1/2
Limit of Detection)
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
Mercury Concentration (ppm)
Spanish Mackerel
182
Q.
3
o
3
110
010
• Average Composite
Sample Cone.
•Maine Action Level
•US EPA Screening
Level
•Florida/EU/Canada
Action Level
•US FDA Action
Level
0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8
0.9
1.0
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-l 5
-------
Mahi-mahi
g
= 019
o
o
4
4
4
4
4
4
O
<
1
:
i
i
i
i
i
i
:
i
j ?
i i
\ *
i i
\ i
\ i
\ i
\ i
\ i
i , >
; ^
(
<
i
i
I
(
(
<
(
(
r
i
i
i
i
i
i lAwpragp Composite
Sample Cone.
-^"Maine Action Level
—^— US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
— ^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
Monkfish
0.8
0.9
1.0
127
070
026
090
9 246
Q.
5 165
218
234
011
053
=
1
1 <
1
4
D <
4
1
> |
> 1
> 1
> 1
> 1
> 1
> |
> 1
s r
| t
\ *
1 ,
\ i
\ i
\ i
\ t
\ t
i , t
(
<
<
(
(
<
(
r
i
1
I
I
1
^^US EPA Screening
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8
0.9
1.0
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-16
-------
Mussel
032
191
175
= 039
o
5
007
042
198
_| 4
\ <
'
1
J <
. <
J <
(ND) <
4
(ND)
/ 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
s r
j ?
1 ,
1 ,
1 i
\ t
t
1 ,
1 ,
1 ,
i , t
: ^
(
<
i
i
I
(
<
<
(
(
r
i
i
i
i
i lAwpragp Composite
Sample Cone.
-^"Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
Oyster
0.8
0.9
1.0
073 |j
041 1
189 1 <
Q 174
Q.
3
O
5 143
-
197
044
190
4
4
4
4
(ND)
4
(ND)
r 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
s r
j j
1 i
\ *
1 ,
\ i
\ i
\ i
\ i
\ i
i , t
-t ^
(
<
<
(
(
<
(
(
r
i
i
1
I
I
1
1
1
i lAvpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8
0.9
1.0
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-17
-------
Ocean Perch
g
= 094
o
(5
4
4
4
4
4
4
4
4
r 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
s r
j ?
1 i
\ *
1 i
\ i
\ i
\ i
\ i
\ i
i , t
: ^
(
<
i
i
I
(
(
<
(
(
r
i
i
i
i
i
i lAwpragp Composite
Sample Cone.
-^"Maine Action Level
—^— US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
— ^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
Pollock
0.8
0.9
1.0
148
-
067
-
186
-
233
Q
= 083
o
O
051
-
259
-
109
-
221
n
4
4
4
4
4
4
4
4
> I
> I
> I
> I
> I
> I
> I
> I
I i
I i
I I
I I
I i
I i
I i
I i
(
<
<
(
(
<
(
(
I
I
1
|
|
1
1
1
i lAvpragp Composite
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8
0.9
1.0
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-l 8
-------
Porgy
m
-
024
-
152
g
Q.
3
o
O
087
125
009
nN
<
4
4
4
4
4
4
4
> I
> I
> I
> I
> I
> I
> I
> I
s r
I i
I ,
I .
I .
I ,
I i
I i
I i
i , t
(
<
<
(
(
<
(
(
r
i
i
1
I
I
1
1
1
i lAvpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
Atlantic Salmon
0.8 0.9
1.0
102
206
135
185
Q
= 178
o
O
059
196
121
034
4
4
4
(ND)
4
(ND)
4
(ND)
4
(ND)
4
(ND)
(ND)
r 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
s r
j ?
1 i
\ *
1 .
1 .
1 i
\ i
\ i
It
i
i , t
: ^
(
<
C
(
(
<
(
«"
r
i
i
•
i
i
i
i
i lAwpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8 0.9
1.0
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-19
-------
Scallop
257 1
f <
232
208
Q
= 194
o
is
066
172
006
(ND) <
(ND)
4
(ND)
4
(ND)
(ND) <
4
(ND)
r 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
s r
j ?
1 i
\ *
1 i
\ i
\ i
I/
i
\ i
\ i
i , t
: ^
(
<
<
(
(
<
<
r
i
i
•
i
i
i
i
i lAwpragp Composite
Sample Cone.
-^"Maine Action Level
—^— US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
— ^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
Shrimp
0.8
0.9
1.0
176 1
f <
134
014
Q
= 056
o
o
057
015
177
(ND) <
(ND)
4
(ND)
4
(ND)
(ND) <
4
(ND)
r 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
s r
j ?
1 i
\ *
1 ,
\ i
\ i
I/
i
\ i
\ i
i , t
-t ^
(
<
<
(
(
<
<
r
i
i
•
i
i
i
i
i lAwpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8
0.9
1.0
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-20
-------
Skate
217
131
052
167
219
068
Q
- 028
1 <
|
1 <
i
ZZ <
ZZ
1 4
s
= -^™
2 151 1 1
O J 1 <
089 •
256 [] <
-tr
255
M <
248 "h
220 n <
238 (ND)
r 1
> 1
> 1
> 1
> |
> 1
> 1
> 1
> 1
v r
j ?
1 i
\ i
\ i
\ i
\ >
\ i
\ i
\ i
-\
-t ^
(
<
i
i
1
{
(
<
(
(
r
i
i
i
i
i
i lAwpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Mercury Concentration (ppm)
Snapper
f ^ /™^
084 | <
086
029
150
188
1
1 <
ZZ
m <
119 | 1
Q -] | <
a 1* tl
Q 106 in
o -J — <
118 ri
f
100 <•
-R
245 |J
251 P <
097 "n
085 | <
149 |
' 1
> 1
> 1
> 1
> |
> 1
> |
> 1
> 1
s r
j i
1 i
\ *
1 ,
\ <
1 i
\ t
\ i
\ i
i , t
, V
(
<
i
i
•
(
(
(
(
(
r
i
i
i
i
i
i lAwpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8
0.9
1.0
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-21
-------
Squid
021
123
140
179
063
- 230
2 254
O "^
253
231
229
078
013
4
4
4
4
4
ND) <
ND)
r 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
> 1
s r
j ?
1 i
\ *
1 i
\ i
\ i
\ i
\ i
\ i
i , t
: ^
(
<
<
(
(
<
(
(
r
i
i
1
I
I
1
1
1
i iA\/pragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
Swordfish
205
161
o
o
035
018
0.1
0.8
0.9
1.0
0.2
0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8
0.9
1.0
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-22
-------
Tilapia
243 |~~|
241 r~i
168 [] <
003 J >
055 1
9 M <
= 227
o
O
228
226
139
117
045
(ND)
4
(ND)
(ND) <
(ND) <
(ND)
4
(ND)
r 1
> \
> \
> \
> \
> \
> \
> \
> \
s r
j ?
1 i
\ *
1 i
\ i
\ i
\ i
\ i
\ i
i , t
: ^
(
<
C
(
(
<
(
(
r
i
i
1
I
I
1
1
1
i lAvpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
Rainbow Trout
0.8
0.9
1.0
g
= 095
o
o
^
4
4
4
4
4
4
4
r 1
> \
> \
> \
> \
> \
> \
> \
> \
s r
j ?
1 i
\ *
1 ,
\ i
\ i
\ i
\ i
\ i
i , t
-t ^
(
<
i
i
I
(
(
<
(
(
r
i
i
i
i
i
i lAwpragp Composite
Sample Cone.
^^Maine Action Level
^^US EPA Screening
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Mercury Concentration (ppm)
Figure A-l. Measured average mercury concentrations by composite sample
vs. selected state and federal action levels, (continued)
A-23
-------
Tuna
Q.
3
O
(5
111
-
201
-
202
-
200
-
158
-
157
-
037
-
036
-
184
-
204
-
183
_
160
<
X
^
N
X
<
I
>
4
> :
v £
> r
' L
^ L~
> L
> I
> I
s r
]
]
1
J
]
J
J
]
i
L
^
\
HI
1
1
i
^
1
<
^
<
(
\
i
<
r
i
•
i
i
i
1 lA^rag'3 l •nmpn=;itfi
Sample Cone.
^^Maine Action Level
Level
^^FIorida/EU/Canada
Action Level
^^US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
Whiting
180
216
113
g 249
a.
3
O
O 141
079
022
008 (ND)
0.8
0.9
1.0
lAwpragp Composite
Sample Cone.
Maine Action Level
US EPA Screening
Level
FIorida/EU/Canada
Action Level
US FDA Action
Level
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Mercury Concentration (ppm)
0.8
0.9
1.0
Figure A-l. Measured average mercury concentrations by composite
sample vs. selected state and federal action levels, (continued)
A-24
-------
Table A-4. Information about each composite sample group used in the analysis
Sam.
Grp.
1
2
3
4
6
7
8
9
10
11
12
13
14
15
17
18
19
Market
Name
Flounder/
Fluke/ Sole
Flounder/
Fluke/ Sole
Tilapia
Cod
Scallop
Mussel
Whiting
Porgy
Spanish
Mackerel
Monkfish
Sea Bass
Squid
Shrimp
Shrimp
Tuna
Swordfish
Mahi-mahi
Common
Name
Flounder,
Blackback
Flounder,
Yellowtail
Tilapia
Cod, Atlantic
Scallop, Sea
Mussel, Blue
Whiting/ Silver
Hake
Porgy/ Scup
Spanish
Mackerel
Monkfish/
Goosefish
Bass, Black
Sea
Squid
Shrimp, White
Shrimp, Black
Tiger
Tuna, Bigeye
Swordfish
Dolphin/ Mahi-
mahi
Whole or
Partial
Partial
Partial
Partial
Partial
Whole
Whole
Whole
Whole
Whole
Partial
Whole
Partial
Partial
Partial
Partial
Partial
Partial
Form of
Sample
fillet
fillet
fillet
fillet
shellfish,
shelled
shellfish
whole
whole
whole
headed and
gutted
whole
tubes U5
headed
headed
fillet / loin
loin
gutted
Water Body
Notes
Atlantic off
Nova Scotia
Atlantic
Farm
Atlantic:
Maine
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Pacific
Farm (Pacific)
Farm
Pacific
Atlantic
Atlantic
Wild or
Farmed
Wild
Wild
Farmed
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Farmed
Farmed
Wild
Wild
Wild
Water Body
(Uni-form)
Atlantic, Wild
Atlantic, Wild
Farm
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Pacific, Wild
Farm
Farm
Pacific, Wild
Atlantic, Wild
Atlantic, Wild
Avg. Total
Length (cm)
48.3
27.0
36.3
62.0
49.3
36.3
23.7
110.0
Avg. Caudal
Length (cm)
23.0
29.0
51.7
37.7
29.0
Avg. Whole
FishWgt. (g)
783.8
746.5
142.6
875.7
1,475.9
1,576.3
849.2
219.2
Num. in
Comp.
13
16
6
3
18
155
12
3
3
3
3
3
75
40
3
3
2
Weight of
Comp. (g)
669.7
689
1,084.8
970
818.9
608.2
368.6
530.8
595.7
596.3
358.6
370.7
603.8
616.3
624.6
723.5
408.6
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
20
21
22
23
24
25
26
28
29
30
31
32
33
34
35
36
37
39
Market
Name
Mackerel
Squid
Whiting
Croaker
Porgy
Sea Bass
Monkfish
Skate
Snapper
Flounder/
Fluke/ Sole
Flounder/
Fluke/ Sole
Mussel
Clam
Atlantic
Salmon
Swordfish
Tuna
Tuna
Mussel
Common
Name
Mackerel,
Atlantic
Squid, Longfin
Whiting/ Silver
Hake
Croaker,
Atlantic
Porgy/ Scup
Bass, Black
Sea
Monkfish/
Goosefish
Skate, Winter
Snapper, Red
Flounder,
Blackback
Sole, Gray
Mussel, Blue
Quahog,
Northern/ Little
Neck Clam
Atlantic
Salmon
Swordfish
Tuna,
Yellowfm
Tuna, Bigeye
Mussel, Blue
Whole or
Partial
Whole
Whole
Whole
Whole
Whole
Whole
Partial
Partial
Whole
Whole
Whole
Whole
Whole
Whole
Partial
Partial
Partial
Whole
Form of
Sample
whole
whole
whole
whole
whole
whole
headed
wings
whole,
gutted
whole
whole
shellfish
shellfish
whole,
gutted
loin
loin
loin
shellfish
Water Body
Notes
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Gulf
Atlantic
Atlantic
Atlantic
Sound
Atlantic
Pacific
Pacific
Pacific
Atlantic
Wild or
Farmed
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Water Body
(Uni-form)
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Sound
Atlantic, Wild
Pacific, Wild
Pacific, Wild
Pacific, Wild
Atlantic, Wild
Avg. Total
Length (cm)
31.5
38.7
28.6
32.3
32.7
33.7
48.0
38.5
34.3
33.0
41.3
66.0
Avg. Caudal
Length (cm)
27.3
24.1
26.8
26.3
27.0
36.0
27.7
27.0
34.7
55.7
Avg. Whole
FishWgt. (g)
248.4
113.6
164.4
453.8
658.6
635.6
1,569.1
908.3
658.0
473.8
474.6
746.5
2,549.9
39,812.3
31,725.4
20,528.2
963.8
Num. in
Comp.
6
12
8
6
3
3
1
2
3
3
3
66
19
3
3
2
40
45
Weight of
Comp. (g)
357.2
695.5
272.8
481.1
271.9
257.7
960.2
565.8
418.1
209.2
207.1
164.2
178.4
649.1
646.3
433.5
616.3
333.9
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
40
41
42
43
44
45
46
47
48
50
51
52
53
54
55
56
Market
Name
Clam
Oyster
Mussel
Clam
Oyster
Tilapia
Flounder/
Fluke/ Sole
Flounder/
Fluke/ Sole
Flounder/
Fluke/ Sole
Cod
Pollock
Skate
Monkfish
Sea Bass
Tilapia
Shrimp
Common
Name
Quahog,
Northern/
Cherrystone
Clam
Oyster, Eastern
Mussel, Blue
Quahog,
Northern/ Little
Neck Clam
Oyster, Eastern
Tilapia
Sole, Gray
Flounder,
Summer
Flounder,
Blackback
Cod, Atlantic
Pollock
Skate, Winter
Monkfish/
Goosefish
Bass, Black
Sea
Tilapia
Shrimp, White
Whole or
Partial
Whole
Whole
Whole
Whole
Whole
Partial
Whole
Whole
Whole
Whole
Partial
Partial
Partial
Whole
Partial
Partial
Form of
Sample
shellfish
shellfish
shellfish
shellfish
shellfish
fillet
whole
whole
whole
Whole,
gutted
headed and
gutted
wings
tails
whole
(round)
fillet
headed
Water Body
Notes
Atlantic
Long Island
Sound, Conn.
Side
Atlantic
Atlantic
Galveston
Bay
Farm
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Atlantic
Farm
Farm (Pacific)
Wild or
Farmed
Wild
Wild
Wild
Wild
Farmed
Farmed
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Farmed
Farmed
Water Body
(Uni-form)
Atlantic, Wild
Sound
Atlantic, Wild
Atlantic, Wild
Farm
Farm
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Farm
Farm
Avg. Total
Length (cm)
41.7
47.0
35.7
62.0
70.0
30.7
40.3
33.7
Avg. Caudal
Length (cm)
35.0
39.7
27.3
53.0
56.0
30.7
25.7
Avg. Whole
FishWgt. (g)
889.5
497.3
1,283.2
553.0
1,954.9
3,639.0
699.8
846.0
557.5
Num. in
Comp.
11
9
47
21
22
6
3
3
3
3
3
3
3
3
3
40
Weight of
Comp. (g)
328.7
150.2
286.7
159
277.5
618
160.4
648.3
360
600.8
601.1
615.9
634.8
318.6
1,084.8
605.1
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
57
58
59
61
62
63
64
65
66
67
68
69
70
72
73
74
Market
Name
Shrimp
Tuna
Atlantic
Salmon
Flounder/
Fluke/ Sole
Cod
Squid
Clam
Blue Crab
Scallop
Pollock
Skate
Chilean Sea
Bass
Monkfish
Porgy
Oyster
Halibut
Common
Name
Shrimp, White
Tuna,
Yellowfm
Atlantic
Salmon
Flounder,
Summer
Cod, Atlantic
Squid, Longfin
Quahog,
Northern/ Little
Neck Clam
Blue Crab/
Softshell
Scallop, Sea
Pollock
Skate, Winter
Chilean Sea
Bass
Monkfish/
Goosefish
Porgy/ Scup
Oyster, Eastern
Halibut, Pacific
Whole or
Partial
Partial
Partial
Partial
Whole
Partial
Partial
Whole
Whole
Whole
Partial
Partial
Whole
Partial
Whole
Whole
Partial
Form of
Sample
headed and
deveined
loin
fillet
whole
headed and
gutted
whole
shellfish
softshell
shelled, dry
headed and
gutted
wings
steak
(headed
and gutted)
Tail
(headed
and gutted)
whole
shellfish
headed and
gutted
Water Body
Notes
Farm
Pacific
N. Pacific
N. Atlantic
N. Atlantic
N. Atlantic
Farm
N. Atlantic
N. Atlantic
N. Atlantic
N. Atlantic
Pacific
N. Atlantic
N. Atlantic
Farm
N. Pacific
Wild or
Farmed
Farmed
Wild
Wild
Wild
Wild
Wild
Farmed,
Cherry
Stone
Creek
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Farmed
(Running
Channel)
Wild
Water Body
(Uni-form)
Farm
Pacific, Wild
Pacific, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Farm
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Pacific, Wild
Atlantic, Wild
Atlantic, Wild
Farm
Pacific, Wild
Avg. Total
Length (cm)
48.7
62.7
32.5
65.7
36.7
17.0
53.3
37.0
71.5
Avg. Caudal
Length (cm)
41.3
29.7
Avg. Whole
FishWgt. (g)
1,424.1
3,664.6
43.3
1,373.9
922.0
3,580.1
837.7
392.9
2,090.0
819.7
2,207.5
45,17.1
Num. in
Comp.
89
3
3
3
3
13
33
11
18
3
3
3
3
3
34
2
Weight of
Comp. (g)
623.4
799.6
649.1
610.8
631.7
312
216
786
864.9
626.2
605.4
624.4
634.1
495.8
177.1
436.4
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
75
76
77
78
79
80
81
83
84
85
86
87
88
89
90
91
92
Market
Name
Sea Bass
Flounder/
Fluke/ Sole
Cod
Squid
Whiting
Catfish
Blue Crab
Pollock
Snapper
Snapper
Snapper
Porgy
Herring
Skate
Monkfish
Croaker
Bluefish
Common
Name
Bass, Black
Sea
Sole, Gray
Cod, Atlantic
Squid,
Japanese
Flying
Whiting/ Silver
Hake
Catfish,
Channel
Blue Crab/
Hardshell
Pollock
Snapper,
Yellowtail
Snapper,
Vermilion
Snapper, Red
Porgy/ Scup
Herring,
Atlantic
Skate, Winter
Monkfish/
Goosefish
Croaker,
Atlantic
Bluefish
Whole or
Partial
Whole
Whole
Partial
Partial
Whole
Whole
Whole
Partial
Partial
Partial
Partial
Partial
Whole
Partial
Partial
Whole
Whole
Form of
Sample
whole
whole
headed and
gutted
tube
whole
whole
hardshell
headed and
gutted
gutted
gutted
gutted
gutted
whole
wings
tail (headed
and gutted)
whole
whole
Water Body
Notes
N. Atlantic
N. Atlantic
N. Atlantic
Pacific
N. Atlantic
Great Lakes
N. Atlantic
N. Atlantic
S. Atlantic
S. Atlantic
S. Atlantic
N. Atlantic
N. Atlantic
N. Atlantic
N. Atlantic
N. Atlantic
N. Atlantic
Wild or
Farmed
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Water Body
(Uni-form)
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Pacific, Wild
Atlantic, Wild
Lake
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Avg. Total
Length (cm)
52.3
40.7
62.0
28.0
43.3
20.0
68.3
38.8
35.1
37.6
31.4
17.0
42.8
50.7
38.0
68.7
Avg. Caudal
Length (cm)
41.0
33.7
35.5
28.0
27.5
29.1
25.4
13.9
31.3
55.0
Avg. Whole
FishWgt. (g)
2,007.0
517.9
6,179.7
601.4
145.6
751.2
4,101.4
3,439.8
521.1
486.6
677.7
549.3
54.2
883.0
1,513.3
701.5
2,893.4
Num. in
Comp.
3
3
2
5
10
3
21
3
3
3
3
3
11
3
3
3
3
Weight of
Comp. (g)
634
250.8
616.4
543.3
371.1
349.3
601.8
639
324.5
455.4
472.7
261.9
128.6
591.4
602.2
300.9
597.7
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
94
95
96
97
98
99
100
101
102
103
105
106
107
108
109
110
Market
Name
Ocean Perch
Rainbow
Trout
Sea Bass
Snapper
Bass
Sea Bass
Snapper
Catfish
Atlantic
Salmon
Flounder/
Fluke/ Sole
Croaker
Snapper
Snapper
Catfish
Pollock
Spanish
Mackerel
Common
Name
Ocean Perch
Rainbow Trout
Bass, Black
Sea
Snapper, Lane
Bass, Hybrid
Striped
Bass, Black
Sea
Snapper,
Yellowtail
Catfish,
Channel
Atlantic
Salmon
Sole, Gray
Croaker,
Atlantic
Snapper,
Yellowtail
Snapper,
Vermilion
Catfish, White
Pollock
Spanish
Mackerel
Whole or
Partial
Whole
Partial
Whole
Partial
Whole
Whole
Partial
Whole
Partial
Whole
Whole
Whole
Whole
Whole
Partial
Whole
Form of
Sample
fillet
gutted
whole
gutted
whole
whole
gutted
whole
gutted
whole
whole
whole
whole
whole
headed &
gutted
whole
Water Body
Notes
China (for
filleting, N.
Atlantic fish)
Farm
N. Atlantic
Pacific
Farm
N. Atlantic
N. Atlantic
Unknown
Atlantic
N. Atlantic
N. Atlantic
S. Atlantic
S. Atlantic
Farm
N. Atlantic
N. Atlantic
Wild or
Farmed
Wild
Farmed
Wild
Wild
Farmed
Wild
Wild
Wild
Farmed
Wild
Wild
Wild
Wild
Farmed
Wild
Wild
Water Body
(Uni-form)
Atlantic, Wild
Farm
Atlantic, Wild
Pacific, Wild
Farm
Atlantic, Wild
Atlantic, Wild
Unknown
Farm
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Farm
Atlantic, Wild
Atlantic, Wild
Avg. Total
Length (cm)
16.0
36.3
40.7
42.7
30.2
34.3
39.7
44.3
87.7
38.7
32.0
36.0
32.3
36.7
68.0
42.0
Avg. Caudal
Length (cm)
32.0
32.0
34.7
24.5
26.7
28.3
39.3
74.3
32.0
27.3
26.7
25.3
31.0
35.0
Avg. Whole
FishWgt. (g)
61.0
571.5
1,005.6
1,075.2
396.7
558.6
481.3
979.7
7,565.7
424.1
454.0
527.3
451.7
795.0
3,344.1
422.4
Num. in
Comp.
10
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
Weight of
Comp. (g)
608.8
574.4
597.3
615.8
201.1
284.2
417.9
517.3
634.3
375.6
371.9
507.8
396.6
487.2
604.4
401
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
Ill
112
113
114
116
117
118
119
120
121
122
123
124
125
127
128
129
130
131
Market
Name
Tuna
Flounder/
Fluke/ Sole
Whiting
Catfish
Croaker
Tilapia
Snapper
Snapper
Sea Bass
Atlantic
Salmon
Flounder/
Fluke/ Sole
Squid
Catfish
Porgy
Monkiish
Croaker
Sea Bass
Bluefish
Skate
Common
Name
Tuna, Bigeye
Flounder,
Summer
Whiting/ Silver
Hake
Catfish, Blue
Croaker,
Atlantic
Tilapia
Snapper,
Caribbean Red
Snapper, Lane
Bass, Black
Sea
Atlantic
Salmon
Flounder,
Summer
Squid, Longfin
Catfish,
Channel
Porgy/ Scup
Monkfish/
Goosefish
Croaker,
Atlantic
Bass, Black
Sea
Bluefish
Skate, Winter
Whole or
Partial
Partial
Whole
Whole
Whole
Whole
Partial
Partial
Partial
Whole
Partial
Whole
Whole
Whole
Whole
Partial
Whole
Whole
Whole
Partial
Form of
Sample
loin
whole
whole
whole
whole
fillet
gutted
gutted
whole
gutted
whole
whole
whole
whole
tail (headed
and gutted)
whole
whole
whole
wings
Water Body
Notes
Pacific or
Indian
N. Atlantic
N. Atlantic
Farm
N. Atlantic
Unknown
S. Atlantic
Pacific
N. Atlantic
N. Pacific
N. Atlantic
N. Atlantic
Unknown
N. Atlantic
Hampton
Bays
N. Atlantic
N. Atlantic
N. Atlantic
N. Atlantic
Wild or
Farmed
Wild
Wild
Wild
Farmed
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Water Body
(Uni-form)
Pacific, Wild
Atlantic, Wild
Atlantic, Wild
Farm
Atlantic, Wild
Unknown
Atlantic, Wild
Pacific, Wild
Atlantic, Wild
Pacific, Wild
Atlantic, Wild
Atlantic, Wild
Unknown
Atlantic, Wild
Bay
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Avg. Total
Length (cm)
42.0
28.0
44.0
34.3
40.0
43.0
42.3
69.7
43.7
45.0
50.3
29.0
51.0
33.3
33.3
49.3
32.0
Avg. Caudal
Length (cm)
36.0
24.0
37.3
28.3
32.0
35.3
34.0
60.7
37.7
42.7
23.7
28.3
27.3
42.3
Avg. Whole
FishWgt. (g)
71,289.0
795.2
139.9
888.8
557.4
833.6
1,138.8
1,236.8
3,403.1
922.4
162.6
1,595.4
386.6
1,536.2
543.7
561.9
1,007.1
735.3
Num. in
Comp.
3
3
10
3
3
3
3
3
3
3
3
8
3
3
3
3
3
3
3
Weight of
Comp. (g)
389.6
615
407.8
544.5
308.2
476.8
657.6
618.2
653.3
601.3
655.5
679
614.5
282.6
597.2
311.9
313
621.2
608.8
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
132
133
134
135
136
138
139
140
141
142
143
144
145
146
148
149
150
151
152
Market
Name
Bass
Blue Crab
Shrimp
Atlantic
Salmon
Flounder/
Fluke/ Sole
Flounder/
Fluke/ Sole
Tilapia
Squid
Whiting
Catfish
Oyster
Blue Crab
Blue Crab
Lobster
Pollock
Snapper
Snapper
Skate
Porgy
Common
Name
Bass, Striped
Blue Crab/
Hardshell
Shrimp, White
Atlantic
Salmon
Flounder,
Summer
Flounder,
Blackback
Tilapia
Squid, Longfin
Whiting/ Silver
Hake
Catfish, Blue
Oyster, Eastern
Blue Crab/
Softshell
Blue Crab/
Hardshell
Lobster,
American
Pollock
Snapper,
Vermilion
Snapper, Red
Skate, Winter
Porgy/ Scup
Whole or
Partial
Whole
Whole
Partial
Partial
Whole
Whole
Whole
Whole
Whole
Whole
Whole
Whole
Whole
Whole
Partial
Partial
Partial
Partial
Whole
Form of
Sample
whole
hardshell
headed
fillet
whole
whole
whole
whole
whole
whole
shellfish
softshell
hardshell
shellfish
headed and
gutted
whole
whole
wings
whole
Water Body
Notes
Chesapeake
Bay
N. Atlantic
Farm (Pacific)
Farm (Pacific)
N. Atlantic
N. Atlantic
Farm
Atlantic
N. Atlantic
Delaware
River
Farm
N. Atlantic or
Delaware Bay
N. Atlantic
N. Atlantic
N. Atlantic
N. Atlantic
N. Atlantic
N. Atlantic
N. Atlantic
Wild or
Farmed
Wild
Wild
Farmed
Farmed
Wild
Wild
Farmed
Wild
Wild
Wild
Farmed
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Water Body
(Uni-form)
Bay
Atlantic, Wild
Farm
Farm
Atlantic, Wild
Atlantic, Wild
Farm
Atlantic, Wild
Atlantic, Wild
River
Farm
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Avg. Total
Length (cm)
64.7
17.0
38.0
38.0
33.0
21.0
31.7
44.7
13.0
43.0
60.7
29.7
42.0
41.7
35.7
Avg. Caudal
Length (cm)
55.7
29.7
31.7
27.7
27.7
38.3
24.0
35.0
30.0
Avg. Whole
FishWgt. (g)
2,644.4
591.0
871.4
861.3
90.4
288.5
905.7
8,137.3
81.5
614.5
2,605.1
333.4
1,123.9
754.2
807.3
Num. in
Comp.
3
19
37
3
3
3
3
15
3
3
46
11
18
3
3
3
3
3
3
Weight of
Comp. (g)
622.4
616.4
603
622.4
478.6
615.2
604.4
659.5
213.7
605.3
587
903.1
579.6
468.2
628.9
314.4
607.7
597.5
375.5
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
153
154
155
156
157
158
160
161
162
163
164
165
166
167
168
169
Market
Name
Croaker
Bluefish
Sea Bass
Snapper
Tuna
Tuna
Tuna
Swordfish
Flounder/
Fluke/ Sole
Cod
Sea Bass
Monkfish
Catfish
Skate
Tilapia
Blue Crab
Common
Name
Croaker,
Atlantic
Bluefish
Bass, Black
Sea
Snapper,
Yellowtail
Tuna, Bigeye
Tuna, Bigeye
Tuna,
Yellowfm
Swordfish
Flounder,
Blackback
Cod, Atlantic
Bass, Black
Sea
Monkfish/
Goosefish
Catfish,
Channel
Skate, Winter
Tilapia
Blue Crab/
Hardshell
Whole or
Partial
Whole
Whole
Whole
Partial
Partial
Partial
Partial
Partial
Whole
Partial
Whole
Partial
Whole
Partial
Partial
Whole
Form of
Sample
whole
whole
whole
whole
loin
loin
loin
loin
whole
gutted
whole
tail (headed
and gutted)
whole
wings
fillets
hardshell
Water Body
Notes
N. Atlantic
N. Atlantic
Long Island
Sound
N. Pacific or
Gulf of
Mexico
Pacific,
western
Pacific,
western
Pacific,
western
Atlantic,
North
Bay of Fundy
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Farm
Atlantic,
North
Wild or
Farmed
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Farmed
Wild
Water Body
(Uni-form)
Atlantic, Wild
Atlantic, Wild
Sound
Unknown
Pacific, Wild
Pacific, Wild
Pacific, Wild
Atlantic, Wild
Bay
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Farm
Atlantic, Wild
Avg. Total
Length (cm)
32.3
59.0
35.7
26.0
36.0
63.7
33.0
39.3
46.7
40.0
Avg. Caudal
Length (cm)
28.0
51.0
30.0
19.0
30.3
55.3
27.0
40.7
Avg. Whole
FishWgt. (g)
446.4
1,969.0
696.9
203.3
638.7
2,712.7
532.0
762.0
1,022.5
722.3
106.1
Num. in
Comp.
3
3
3
4
2
2
2
3
3
3
3
3
3
3
3
29
Weight of
Comp. (g)
301.7
611.4
507.6
325
555.9
566.1
318.5
686.4
316.9
605.2
356.6
609.8
467.3
517.9
576
640
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
171
172
173
174
175
176
177
178
179
180
182
183
184
185
186
187
Market
Name
Blue Crab
Scallop
Clam
Oyster
Mussel
Shrimp
Shrimp
Atlantic
Salmon
Squid
Whiting
Spanish
Mackerel
Tuna
Tuna
Atlantic
Salmon
Pollock
Cod
Common
Name
Blue Crab/
Softshell
Scallop, Sea
Quahog,
Northern/ Little
Neck Clam
Oyster, Pacific
Mussel, Blue
Shrimp, White
Shrimp, Black
Tiger
Atlantic
Salmon
Squid, Longfin
Whiting,
Offshore
Spanish
Mackerel
Tuna, Bigeye
Tuna, Bigeye
Atlantic
Salmon
Pollock
Cod, Atlantic
Whole or
Partial
Whole
Whole
Whole
Whole
Whole
Partial
Partial
Partial
Whole
Whole
Whole
Partial
Partial
Partial
Partial
Partial
Form of
Sample
softshell
shelled, dry
shellfish
shellfish,
shelled
shellfish
headed
headed
gutted
whole
whole
whole
loin, from
belly
loin, from
belly
gutted
fillet
fillet
Water Body
Notes
Atlantic,
North
Atlantic,
North
Atlantic,
North
Farm
Farm
Farm
Farm
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Pacific
Pacific
West Coast
farm
Atlantic,
North
Atlantic,
North
Wild or
Farmed
Wild
Wild
Wild
Farmed
Farmed
Farmed
Farmed
Farmed
Wild
Wild
Wild
Wild
Wild
Farmed
Wild
Wild
Water Body
(Uni-form)
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Farm
Farm
Farm
Farm
Farm
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Pacific, Wild
Pacific, Wild
Farm
Atlantic, Wild
Atlantic, Wild
Avg. Total
Length (cm)
72.7
23.0
47.0
47.3
33.3
26.3
111
32.5
32.0
Avg. Caudal
Length (cm)
66.7
43.3
39.0
69.3
Avg. Whole
FishWgt. (g)
112.3
3,978.6
69.4
828.8
580.5
530.2
787.5
5,790.7
870.3
432.6
Num. in
Comp.
5
10
50
29
206
38
39
3
14
3
3
3
3
3
2
2
Weight of
Comp. (g)
649.5
603.6
608.9
604.1
600.4
608.2
600.2
607.2
640.2
585.3
535.1
598.9
590.8
604.6
404.8
446.2
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
188
189
190
191
193
194
195
196
197
198
199
200
201
202
204
Market
Name
Snapper
Oyster
Oyster
Mussel
Clam
Scallop
Bass
Atlantic
Salmon
Oyster
Mussel
Clam
Tuna
Tuna
Tuna
Tuna
Common
Name
Snapper, Red
Oyster, Eastern
Oyster, Eastern
Mussel, Blue
Quahog,
Northern/ Little
Neck Clam
Scallop, Sea
Bass, Hybrid
Striped
Atlantic
Salmon
Oyster, Eastern
Mussel, Blue
Quahog,
Northern/ Top
Neck Clam
Tuna,
Yellowfm
Tuna,
Yellowfm
Tuna,
Yellowfin
Tuna,
Yellowiin
Whole or
Partial
Partial
Whole
Whole
Whole
Whole
Whole
Whole
Partial
Whole
Whole
Whole
Partial
Partial
Partial
Partial
Form of
Sample
gutted
shellfish
shellfish
shellfish
shellfish
shelled, dry
whole
fillet
shellfish
shellfish
shellfish
loin
loin
loin
loin
Water Body
Notes
Atlantic,
South
Farm,
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Kent Farm
Farmed
Delaware
Bay, Farmed
Atlantic,
North, Farm
Atlantic,
North, Farm
Pacific, North
Pacific
Pacific, North
Pacific, North
Wild or
Farmed
Wild
Farmed
Wild
Wild
Wild
Wild
Farmed
Farmed
Farmed
Farmed
Farmed
Wild
Wild
Wild
Wild
Water Body
(Uni-form)
Atlantic, Wild
Farm
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Farm
Farm
Farm
Farm
Farm
Pacific, Wild
Pacific, Wild
Pacific, Wild
Pacific, Wild
Avg. Total
Length (cm)
35.3
34.7
40.0
13.3
12.5
13.0
13.3
Avg. Caudal
Length (cm)
29.3
29.0
Avg. Whole
FishWgt. (g)
573.8
2,448.4
2,694.5
1,799.8
1,022.7
871.1
732.1
742.4
3,919.7
826.3
2,906.2
495.3
529.0
547.7
495.0
Num. in
Comp.
3
23
17
153
44
12
3
3
32
44
33
3
2
3
3
Weight of
Comp. (g)
573.1
407.3
238.4
521.9
616.4
594.2
148.4
611.1
565
289
564.2
612.1
604.9
621.8
625.5
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
205
206
207
208
209
210
211
212
213
215
216
217
218
219
220
221
222
Market
Name
Swordfish
Atlantic
Salmon
Blue Crab
Scallop
Croaker
Mackerel
Mackerel
Mackerel
Croaker
Croaker
Whiting
Skate
Monkfish
Skate
Skate
Pollock
Cod
Common
Name
Swordfish
Atlantic
Salmon
Blue Crab/
Softshell
Scallop, Sea
Croaker,
Atlantic
Mackerel,
Atlantic
Mackerel,
Atlantic
Mackerel,
Atlantic
Croaker,
Atlantic
Croaker,
Atlantic
Whiting/ Silver
Hake
Skate, Winter
Monkfish/
Goosefish
Skate, Winter
Skate, Winter
Pollock
Cod, Pacific
Whole or
Partial
Partial
Partial
Whole
Whole
Whole
Whole
Whole
Whole
Whole
Whole
Whole
Partial
Partial
Partial
Partial
Partial
Partial
Form of
Sample
loin
fillet, on
site
softshell
shelled, dry
whole
whole
whole
whole
whole
whole
whole
wings
tail (headed
and gutted)
wings
wings, skin
off
fillet
fillet
Water Body
Notes
Pacific
Farm
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Pacific, North
Wild or
Farmed
Wild
Farm
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Water Body
(Uni-form)
Pacific, Wild
Farm
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Pacific, Wild
Avg. Total
Length (cm)
15.0
48.7
35.0
33.5
33.5
32.0
31.7
34.7
27.0
37.7
37.7
37.7
30.7
49.3
34.3
Avg. Caudal
Length (cm)
30.3
29.5
29.8
28.0
27.0
30.3
Avg. Whole
FishWgt. (g)
442.9
1,513.9
66.3
855.5
566.6
333.6
334.0
368.1
417.5
526.8
140.2
445.1
781.7
379.3
631.7
1,573.4
740.0
Num. in
Comp.
3
3
8
32
3
4
4
4
3
3
10
3
3
3
3
3
3
Weight of
Comp. (g)
610
606.5
654.4
585.3
319.7
299.3
325.4
278.8
240.2
358.6
446.7
325.9
602.2
255.8
614.7
603.6
630.7
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
223
224
226
227
228
229
230
231
232
233
234
235
237
238
239
240
241
Market
Name
Cod
Cod
Tilapia
Tilapia
Tilapia
Squid
Squid
Squid
Scallop
Pollock
Monkfish
Mackerel
Mackerel
Skate
Mackerel
Mackerel
Tilapia
Common
Name
Cod, Pacific
Cod, Pacific
Tilapia
Tilapia
Tilapia
Squid,
Japanese
Flying
Squid,
Japanese
Flying
Squid,
Japanese
Flying
Scallop, Sea
Pollock
Monkfish/
Goosefish
Mackerel,
Atlantic
Mackerel,
Atlantic
Skate, Winter
Mackerel,
Atlantic
Mackerel,
Atlantic
Tilapia
Whole or
Partial
Partial
Partial
Partial
Partial
Partial
Partial
Partial
Partial
Whole
Partial
Partial
Whole
Whole
Partial
Whole
Whole
Partial
Form of
Sample
fillet
fillet
fillet
fillet
fillet
tubes
tubes
tubes
shelled, dry
headed and
gutted
tail (headed
and gutted)
whole
whole
wings
whole
whole
fillet
Water Body
Notes
Pacific, North
Pacific, North
Farm
Farm
Farm
Pacific, North
Pacific, North
Pacific, North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Farm
Wild or
Farmed
Wild
Wild
Farmed
Farmed
Farmed
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Farmed
Water Body
(Uni-form)
Pacific, Wild
Pacific, Wild
Farm
Farm
Farm
Pacific, Wild
Pacific, Wild
Pacific, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Farm
Avg. Total
Length (cm)
34.0
38.3
22.3
23.7
22.7
17.0
17.0
17.0
3.0
64.7
36.3
29.0
31.0
31.7
30.0
29.0
19.3
Avg. Caudal
Length (cm)
26.0
28.0
Avg. Whole
FishWgt. (g)
802.9
902.1
324.6
343.0
361.1
103.9
103.6
94.2
839.3
3,302.8
701.2
202.4
239.7
743.9
216.6
215.0
209.2
Num. in
Comp.
3
3
3
3
3
5
5
5
27
3
3
5
5
3
5
6
3
Weight of
Comp. (g)
603.7
628.4
604.1
659
621.9
428.7
452.1
421.3
616
632.4
614.8
337.7
342.4
596.2
411.2
511.7
503.4
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
242
243
244
245
246
248
249
250
251
252
253
254
255
256
257
259
Market
Name
Blue Crab
Tilapia
Blue Crab
Snapper
Monkfish
Skate
Whiting
Sea Bass
Snapper
Cod
Squid
Squid
Skate
Skate
Scallop
Pollock
Common
Name
Blue Crab/
Softshell
Tilapia
Blue Crab/
Softshell
Snapper, Red
Monkfish/
Goosefish
Skate, Winter
Whiting,
Offshore
Bass, Black
Sea
Snapper, Red
Cod, Atlantic
Squid, Longfin
Squid, Longfin
Skate, Winter
Skate, Winter
Scallop, Sea
Pollock
Whole or
Partial
Whole
Partial
Whole
Whole
Partial
Partial
Whole
Whole
Partial
Partial
Whole
Whole
Partial
Partial
Whole
Partial
Form of
Sample
softshell
fillet
softshell
whole
tail (headed
and gutted)
wings
whole
whole
gutted
gutted
whole
whole
wings
wings
shelled, dry
headed and
gutted
Water Body
Notes
Atlantic,
North
Farm
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
South
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Atlantic,
North
Wild or
Farmed
Wild
Farmed
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Wild
Water Body
(Uni-form)
Atlantic, Wild
Farm
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Atlantic, Wild
Avg. Total
Length (cm)
18.7
47.3
37.7
40.3
35.0
39.7
37.0
58.0
18.0
18.0
37.3
38.7
4.0
64.7
Avg. Caudal
Length (cm)
37.7
33.0
30.3
51.7
Avg. Whole
FishWgt. (g)
54.6
206.1
55.8
1,706.7
779.5
1,175.1
271.5
842.2
599.8
1,942.4
41.9
40.0
888.4
769.8
724.8
3,178.3
Num. in
Comp.
9
3
11
3
3
3
5
3
3
3
17
21
3
3
11
3
Weight of
Comp. (g)
390.1
486.6
507.5
600
589.1
601.7
450.4
575.2
575.9
646.4
448.4
529
595.3
600.1
627.4
597.8
-------
Table A-4. Information about each composite sample group used in the analysis (continued)
Sam.
Grp.
260
Market
Name
Blue Crab
Common
Name
Blue Crab/
Hardshell
Whole or
Partial
Whole
Form of
Sample
hardshell
Water Body
Notes
Atlantic,
North
Wild or
Farmed
Wild
Water Body
(Uni-form)
Atlantic, Wild
Avg. Total
Length (cm)
Avg. Caudal
Length (cm)
Avg. Whole
FishWgt. (g)
116.5
Num. in
Comp.
9
Weight of
Comp. (g)
209.4
VO
-------
Table A-5. Mercury concentrations by composite sample number used in analysis (prior to averaging duplicates and
replicates)
Sample
Number
001.C
001.C
001.C
002.C
003.C
004.C
005.C
006.C
007.C
008.C
009.C
010.C
Oll.C
012.C
013.C
014.C
015.C
016.C
017.C
018.C
019.C
020.C
021.C
021.C
Sample
Group
1
1
1
2
3
4
4
6
7
8
9
10
11
12
13
14
15
15
17
18
19
20
21
21
Duplicate Of
004.C
015.C
Replicate ID
LR-1
LR-2
LR-3
LR-1
LR-2
Market Name
Flounder/Fluke/Sole
Flounder/Fluke/Sole
Flounder/Fluke/Sole
Flounder/Fluke/Sole
Tilapia
Cod
Cod
Scallop
Mussel
Whiting
Porgy
Spanish Mackerel
Monkfish
Sea Bass
Squid
Shrimp
Shrimp
Shrimp
Tuna
Swordfish
Mahi-mahi
Mackerel
Squid
Squid
Common Name
Flounder, Blackback
Flounder, Blackback
Flounder, Blackback
Flounder, Yellowtail
Tilapia
Cod, Atlantic
Cod, Atlantic
Scallop, Sea
Mussel, Blue
Whiting/Silver Hake
Porgy/Scup
Spanish Mackerel
Monkiish/Gooseiish
Bass, Black Sea
Squid
Shrimp, White
Shrimp, Black Tiger
Shrimp, Black Tiger
Tuna, Bigeye
Swordfish
Dolphin/Mahi-mahi
Mackerel, Atlantic
Squid, Longfm
Squid, Longfin
Hg Concentration
(mg/kg)
0.02
0.027
0.024
0.014
0.068
0.11
0.064
0.03
0.13
0.14
0.22
0.013
0.026
0.022
Detected
N
N
N
N
Y
Y
Y
N
Y
N
Y
Y
Y
Y
N
N
N
N
Y
Y
Y
Y
Y
Y
Reporting Limit
(mg/kg)
0.0095
0.0095
0.0097
0.0093
0.01
0.01
0.01
0.0086
0.0087
0.0096
0.01
0.01
0.01
0.01
0.0084
0.0096
0.0085
0.0088
0.01
0.01
0.01
0.009
0.0096
0.0088
-------
Table A-5. Mercury concentrations by composite sample number used in analysis (prior to averaging duplicates and
replicates) (continued)
Sample
Number
021.C
022.C
023.C
024.C
025.C
026.C
027.C
028.C
029.C
030.C
031.C
032.C
033.C
034.C
035.C
036.C
037.C
038.C
039.C
040.C
041.C
041.C
042.C
043.C
044.C
Sample
Group
21
22
23
24
25
26
26
28
29
30
31
32
33
34
35
36
37
37
39
40
41
41
42
43
44
Duplicate Of
026.C
037.C
Replicate ID
LR-3
LR-1
LR-2
Market Name
Squid
Whiting
Croaker
Porgy
Sea Bass
Monkfish
Monkfish
Skate
Snapper
Flounder/Fluke/Sole
Flounder/Fluke/Sole
Mussel
Clam
Atlantic Salmon
Swordfish
Tuna
Tuna
Tuna
Mussel
Clam
Oyster
Oyster
Mussel
Clam
Oyster
Common Name
Squid, Longfm
Whiting/Silver Hake
Croaker, Atlantic
Porgy/Scup
Bass, Black Sea
Monkfish/Goosefish
Monkfish/Goosefish
Skate, Winter
Snapper, Red
Flounder, Blackback
Sole, Gray
Mussel, Blue
Quahog, Northern/Little Neck Clam
Atlantic Salmon
Swordfish
Tuna, Yellowfin
Tuna, Bigeye
Tuna, Bigeye
Mussel, Blue
Quahog, Northern/Cherrystone Clam
Oyster, Eastern
Oyster, Eastern
Mussel, Blue
Quahog, Northern/Little Neck Clam
Oyster, Eastern
Hg Concentration
(mg/kg)
0.024
0.015
0.11
0.12
0.09
0.15
0.14
0.067
0.076
0.036
0.065
0.019
0.39
0.37
0.41
0.41
0.014
0.02
0.019
0.02
Detected
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
Reporting Limit
(mg/kg)
0.0085
0.0093
0.01
0.01
0.01
0.01
0.01
0
0.01
0.01
0.01
0.0094
0.0094
0.0086
0.01
0.01
0.01
0.01
0.0088
0.01
0.0097
0.01
0.0088
0.0085
0.0099
-------
Table A-5. Mercury concentrations by composite sample number used in analysis (prior to averaging duplicates and
replicates) (continued)
Sample
Number
045.C
046.C
047.C
048.C
049.C
050.C
051.C
052.C
053.C
054.C
055.C
056.C
057.C
058.C
058.C
058.C
059.C
060.C
061.C
062.C
063.C
064.C
065.C
066.C
067.C
Sample
Group
45
46
47
48
48
50
51
52
53
54
55
56
57
58
58
58
59
59
61
62
63
64
65
66
67
Duplicate Of
048.C
059.C
Replicate ID
LR-1
LR-2
LR-3
Market Name
Tilapia
Flounder/Fluke/Sole
Flounder/Fluke/Sole
Flounder/Fluke/Sole
Flounder/Fluke/Sole
Cod
Pollock
Skate
Monkfish
Sea Bass
Tilapia
Shrimp
Shrimp
Tuna
Tuna
Tuna
Atlantic Salmon
Atlantic Salmon
Flounder/Fluke/Sole
Cod
Squid
Clam
Blue Crab
Scallop
Pollock
Common Name
Tilapia
Sole, Gray
Flounder, Summer
Flounder, Blackback
Flounder, Blackback
Cod, Atlantic
Pollock
Skate, Winter
Monkiish/Gooseiish
Bass, Black Sea
Tilapia
Shrimp, White
Shrimp, White
Tuna, Yellowfin
Tuna, Yellowfin
Tuna, Yellowfin
Atlantic Salmon
Atlantic Salmon
Flounder, Summer
Cod, Atlantic
Squid, Longfin
Quahog, Northern/Little Neck Clam
Blue Crab/Softshell
Scallop, Sea
Pollock
Hg Concentration
(mg/kg)
0.044
0.054
0.026
0.026
0.033
0.13
0.094
0.054
0.075
0.015
0.044
0.041
0.043
0.1
0.049
0.016
0.009
0.029
0.17
Detected
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
Y
Y
Y
N
N
Y
Y
Y
Y
Y
N
Y
Reporting Limit
(mg/kg)
0.0092
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.0099
0.009
0.0087
0.0092
0.0092
0.0091
0.0095
0.0092
0.02
0.02
0.0094
0.0089
0.02
0.0089
0.02
-------
Table A-5. Mercury concentrations by composite sample number used in analysis (prior to averaging duplicates and
replicates) (continued)
Sample
Number
068.C
069.C
070.C
071. C
072.C
073.C
074.C
075.C
076.C
076.C
076.C
077.C
078.C
079.C
080.C
081.C
082.C
083.C
084.C
085.C
086.C
087.C
088.C
089.C
090.C
Sample
Group
68
69
70
70
72
73
74
75
76
76
76
77
78
79
80
81
81
83
84
85
86
87
88
89
90
Duplicate Of
070.C
081.C
Replicate ID
LR-1
LR-2
LR-3
Market Name
Skate
Chilean Sea Bass
Monkfish
Monkfish
Porgy
Oyster
Halibut
Sea Bass
Flounder/Fluke/Sole
Flounder/Fluke/Sole
Flounder/Fluke/Sole
Cod
Squid
Whiting
Catfish
Blue Crab
Blue Crab
Pollock
Snapper
Snapper
Snapper
Porgy
Herring
Skate
Monkfish
Common Name
Skate, Winter
Chilean Sea Bass
Monkfish/Goosefish
Monkfish/Goosefish
Porgy/Scup
Oyster, Eastern
Halibut, Pacific
Bass, Black Sea
Sole, Gray
Sole, Gray
Sole, Gray
Cod, Atlantic
Squid, Japanese Flying
Whiting/Silver Hake
Catfish, Channel
Blue Crab/Hardshell
Blue Crab/Hardshell
Pollock
Snapper, Yellowtail
Snapper, Vermilion
Snapper, Red
Porgy/Scup
Herring, Atlantic
Skate, Winter
Monkfish/Goosefish
Hg Concentration
(mg/kg)
0.071
0.13
0.15
0.15
0.13
0.047
0.15
0.11
0.054
0.06
0.056
0.049
0.016
0.042
0.023
0.021
0.13
0.082
0.02
0.076
0.087
0.022
0.045
0.14
Detected
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Reporting Limit
(mg/kg)
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.0091
0.0085
0.0093
0.02
0.0092
0.0098
0.02
0.0093
0.0089
0.02
0.02
0.0096
0.02
0.02
0.0087
0.02
0.02
-------
Table A-5. Mercury concentrations by composite sample number used in analysis (prior to averaging duplicates and
replicates) (continued)
Sample
Number
091.C
092.C
093.C
094.C
094.C
094.C
095.C
096.C
097.C
098.C
099.C
100.C
101.C
102.C
103.C
104.C
105.C
106.C
107.C
108.C
109.C
110.C
lll.C
lll.C
lll.C
Sample
Group
91
92
92
94
94
94
95
96
97
98
99
100
101
102
103
103
105
106
107
108
109
110
111
111
111
Duplicate Of
092.C
103.C
Replicate ID
LR-1
LR-2
LR-3
LR-1
LR-2
LR-3
Market Name
Croaker
Bluefish
Bluefish
Ocean Perch
Ocean Perch
Ocean Perch
Rainbow Trout
Sea Bass
Snapper
Bass
Sea Bass
Snapper
Catfish
Atlantic Salmon
Flounder/Fluke/Sole
Flounder/Fluke/Sole
Croaker
Snapper
Snapper
Catfish
Pollock
Spanish Mackerel
Tuna
Tuna
Tuna
Common Name
Croaker, Atlantic
Bluefish
Bluefish
Ocean Perch
Ocean Perch
Ocean Perch
Rainbow Trout
Bass, Black Sea
Snapper, Lane
Bass, Hybrid Striped
Bass, Black Sea
Snapper, Yellowtail
Catfish, Channel
Atlantic Salmon
Sole, Gray
Sole, Gray
Croaker, Atlantic
Snapper, Yellowtail
Snapper, Vermilion
Catfish, White
Pollock
Spanish Mackerel
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Hg Concentration
(mg/kg)
0.056
0.16
0.16
0.022
0.024
0.021
0.012
0.094
0.027
0.014
0.079
0.034
0.024
0.019
0.069
0.073
0.068
0.04
0.083
0.041
0.092
0.15
0.85
0.8
0.81
Detected
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Reporting Limit
(mg/kg)
0.02
0.02
0.02
0.0096
0.0088
0.0092
0.0095
0.02
0.0093
0.0097
0.02
0.02
0.0095
0.0099
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.0096
0.0092
0.0091
-------
Table A-5. Mercury concentrations by composite sample number used in analysis (prior to averaging duplicates and
replicates) (continued)
Sample
Number
112.C
113.C
114.C
115.C
116.C
117.C
118.C
119.C
120.C
121.C
122.C
123.C
124.C
125.C
126.C
127.C
128.C
129.C
129.C
129.C
130.C
131.C
132.C
133.C
134.C
Sample
Group
112
113
114
114
116
117
118
119
120
121
122
123
124
125
125
127
128
129
129
129
130
131
132
133
134
Duplicate Of
114.C
125.C
Replicate ID
LR-1
LR-2
LR-3
Market Name
Flounder/Fluke/Sole
Whiting
Catfish
Catfish
Croaker
Tilapia
Snapper
Snapper
Sea Bass
Atlantic Salmon
Flounder/Fluke/Sole
Squid
Catfish
Porgy
Porgy
Monkfish
Croaker
Sea Bass
Sea Bass
Sea Bass
Bluefish
Skate
Bass
Blue Crab
Shrimp
Common Name
Flounder, Summer
Whiting/Silver Hake
Catfish, Blue
Catfish, Blue
Croaker, Atlantic
Tilapia
Snapper, Caribbean Red
Snapper, Lane
Bass, Black Sea
Atlantic Salmon
Flounder, Summer
Squid, Longfin
Catfish, Channel
Porgy/Scup
Porgy/Scup
Monkfish/Goosefish
Croaker, Atlantic
Bass, Black Sea
Bass, Black Sea
Bass, Black Sea
Bluefish
Skate, Winter
Bass, Striped
Blue Crab/Hardshell
Shrimp, White
Hg Concentration
(mg/kg)
0.1
0.031
0.045
0.045
0.13
0.039
0.052
0.083
0.067
0.023
0.032
0.079
0.092
0.18
0.092
0.064
0.066
0.054
0.12
0.11
0.047
0.019
Detected
Y
Y
Y
Y
Y
N
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
Reporting Limit
(mg/kg)
0.02
0.02
0.02
0.02
0.02
0.0097
0.02
0.02
0.02
0.0086
0.02
0.0089
0.02
0.02
0.02
0.02
0.02
0.0086
0.009
0.0093
0.02
0.02
0.02
0.0097
0.0098
-------
Table A-5. Mercury concentrations by composite sample number used in analysis (prior to averaging duplicates and
replicates) (continued)
Sample
Number
135.C
136.C
137.C
138.C
139.C
140.C
141.C
142.C
143.C
144.C
145.C
146.C
147.C
147.C
147.C
148.C
149.C
150.C
151.C
152.C
153.C
154.C
155.C
156.C
157.C
Sample
Group
135
136
136
138
139
140
141
142
143
144
145
146
146
146
146
148
149
150
151
152
153
154
155
156
157
Duplicate Of
136.C
146.C
146.C
146.C
Replicate ID
LR-1
LR-2
LR-3
Market Name
Atlantic Salmon
Flounder/Fluke/Sole
Flounder/Fluke/Sole
Flounder/Fluke/Sole
Tilapia
Squid
Whiting
Catfish
Oyster
Blue Crab
Blue Crab
Lobster
Lobster
Lobster
Lobster
Pollock
Snapper
Snapper
Skate
Porgy
Croaker
Bluefish
Sea Bass
Snapper
Tuna
Common Name
Atlantic Salmon
Flounder, Summer
Flounder, Summer
Flounder, Blackback
Tilapia
Squid, Longfm
Whiting/Silver Hake
Catfish, Blue
Oyster, Eastern
Blue Crab/Softshell
Blue Crab/Hardshell
Lobster, American
Lobster, American
Lobster, American
Lobster, American
Pollock
Snapper, Vermilion
Snapper, Red
Skate, Winter
Porgy/Scup
Croaker, Atlantic
Bluefish
Bass, Black Sea
Snapper, Yellowtail
Tuna, Bigeye
Hg Concentration
(mg/kg)
0.012
0.045
0.044
0.049
0.018
0.02
0.028
0.011
0.017
0.073
0.07
0.069
0.063
0.18
0.017
0.065
0.06
0.096
0.086
0.16
0.067
0.047
0.46
Detected
Y
Y
Y
Y
N
Y
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Reporting Limit
(mg/kg)
0.0095
0.02
0.02
0.02
0.0097
0.0083
0.009
0.0092
0.0081
0.0088
0.0098
0.02
0.0086
0.0092
0.0088
0.02
0.0092
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
-------
Table A-5. Mercury concentrations by composite sample number used in analysis (prior to averaging duplicates and
replicates) (continued)
Sample
Number
158.C
159.C
160.C
161.C
162.C
163.C
164.C
165.C
165.C
165.C
166.C
167.C
168.C
169.C
170.C
171.C
172.C
173.C
174.C
175.C
176.C
177.C
178.C
179.C
180.C
Sample
Group
158
158
160
161
162
163
164
165
165
165
166
167
168
169
169
171
172
173
174
175
176
177
178
179
180
Duplicate Of
158.C
169.C
Replicate ID
LR-1
LR-2
LR-3
Market Name
Tuna
Tuna
Tuna
Swordfish
Flounder/Fluke/Sole
Cod
Sea Bass
Monkfish
Monkfish
Monkfish
Catfish
Skate
Tilapia
Blue Crab
Blue Crab
Blue Crab
Scallop
Clam
Oyster
Mussel
Shrimp
Shrimp
Atlantic Salmon
Squid
Whiting
Common Name
Tuna, Bigeye
Tuna, Bigeye
Tuna, Yellowfm
Swordfish
Flounder, Blackback
Cod, Atlantic
Bass, Black Sea
Monkiish/Gooseiish
Monkiish/Gooseiish
Monkfish/Goosefish
Catfish, Channel
Skate, Winter
Tilapia
Blue Crab/Hardshell
Blue Crab/Hardshell
Blue Crab/Softshell
Scallop, Sea
Quahog, Northern/Little Neck Clam
Oyster, Pacific
Mussel, Blue
Shrimp, White
Shrimp, Black Tiger
Atlantic Salmon
Squid, Longfin
Whiting, Offshore
Hg Concentration
(mg/kg)
0.5
0.49
0.13
0.51
0.04
0.018
0.059
0.086
0.075
0.078
0.094
0.084
0.021
0.024
0.025
0.013
0.015
0.011
0.016
0.077
Detected
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
Y
Y
Y
N
N
Y
Y
Reporting Limit
(mg/kg)
0.02
0.02
0.02
0.02
0.02
0.0091
0.02
0.0084
0.0088
0.0096
0.02
0.02
0.0088
0.0094
0.0088
0.0085
0.0086
0.0088
0.0097
0.0085
0.0087
0.0084
0.0098
0.0101
0.02
-------
Table A-5. Mercury concentrations by composite sample number used in analysis (prior to averaging duplicates and
replicates) (continued)
Sample
Number
181.C
182.C
183.C
183.C
183.C
184.C
185.C
186.C
187.C
188.C
189.C
190.C
191.C
192.C
193.C
194.C
195.C
196.C
197.C
198.C
199.C
200.C
201.C
201.C
201.C
Sample
Group
180
182
183
183
183
184
185
186
187
188
189
190
191
191
193
194
195
196
197
198
199
200
201
201
201
Duplicate Of
180.C
191.C
Replicate ID
LR-1
LR-2
LR-3
LR-1
LR-2
LR-3
Market Name
Whiting
Spanish Mackerel
Tuna
Tuna
Tuna
Tuna
Atlantic Salmon
Pollock
Cod
Snapper
Oyster
Oyster
Mussel
Mussel
Clam
Scallop
Bass
Atlantic Salmon
Oyster
Mussel
Clam
Tuna
Tuna
Tuna
Tuna
Common Name
Whiting, Offshore
Spanish Mackerel
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Atlantic Salmon
Pollock
Cod, Atlantic
Snapper, Red
Oyster, Eastern
Oyster, Eastern
Mussel, Blue
Mussel, Blue
Quahog, Northern/Little Neck Clam
Scallop, Sea
Bass, Hybrid Striped
Atlantic Salmon
Oyster, Eastern
Mussel, Blue
Quahog, Northern/Top Neck Clam
Tuna, Yellowfin
Tuna, Yellowfin
Tuna, Yellowfin
Tuna, Yellowfin
Hg Concentration
(mg/kg)
0.072
0.2
0.18
0.2
0.19
0.36
0.14
0.027
0.06
0.013
0.016
0.016
0.014
0.01
0.01
0.59
0.78
0.82
0.81
Detected
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
Y
N
Y
Y
N
N
Y
N
Y
N
Y
Y
Y
Y
Y
Reporting Limit
(mg/kg)
0.02
0.02
0.0084
0.0094
0.0095
0.02
0.01
0.02
0.0093
0.02
0.0097
0.0092
0.009
0.009
0.0085
0.0094
0.0093
0.0088
0.0091
0.0087
0.0084
0.02
0.0088
0.0091
0.0091
-------
Table A-5. Mercury concentrations by composite sample number used in analysis (prior to averaging duplicates and
replicates) (continued)
Sample
Number
202.C
203.C
204.C
205.C
206.C
207.C
208.C
209.C
210.C
211.C
212.C
213.C
214.C
215.C
216.C
217.C
218.C
219.C
219.C
219.C
220.C
221.C
222.C
223.C
224.C
Sample
Group
202
202
204
205
206
207
208
209
210
211
212
213
213
215
216
217
218
219
219
219
220
221
222
223
224
Duplicate Of
202.C
213.C
Replicate ID
LR-1
LR-2
LR-3
Market Name
Tuna
Tuna
Tuna
Swordfish
Atlantic Salmon
Blue Crab
Scallop
Croaker
Mackerel
Mackerel
Mackerel
Croaker
Croaker
Croaker
Whiting
Skate
Monkfish
Skate
Skate
Skate
Skate
Pollock
Cod
Cod
Cod
Common Name
Tuna, Yellowfm
Tuna, Yellowfin
Tuna, Yellowfin
Swordfish
Atlantic Salmon
Blue Crab/Softshell
Scallop, Sea
Croaker, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Croaker, Atlantic
Croaker, Atlantic
Croaker, Atlantic
Whiting/Silver Hake
Skate, Winter
Monkiish/Gooseiish
Skate, Winter
Skate, Winter
Skate, Winter
Skate, Winter
Pollock
Cod, Pacific
Cod, Pacific
Cod, Pacific
Hg Concentration
(mg/kg)
0.7
0.66
0.35
0.57
0.014
0.077
0.034
0.03
0.03
0.057
0.058
0.082
0.033
0.12
0.078
0.083
0.074
0.064
0.022
0.079
0.04
0.024
0.024
Detected
Y
Y
Y
Y
Y
N
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Reporting Limit
(mg/kg)
0.02
0.02
0.02
0.02
0.0088
0.01
0.0098
0.0093
0.0094
0.0092
0.0095
0.02
0.02
0.02
0.0099
0.02
0.02
0.0088
0.0095
0.0089
0.0089
0.02
0.02
0.0095
0.0096
-------
Table A-5. Mercury concentrations by composite sample number used in analysis (prior to averaging duplicates and
replicates) (continued)
Sample
Number
225.C
226.C
227.C
228.C
229.C
230.C
231.C
232.C
233.C
234.C
235.C
236.C
237.C
237.C
237.C
238.C
239.C
240.C
241.C
242.C
243.C
244.C
245.C
246.C
247.C
Sample
Group
224
226
227
228
229
230
231
232
233
234
235
235
237
237
237
238
239
240
241
242
243
244
245
246
246
Duplicate Of
224.C
235.C
246.C
Replicate ID
LR-1
LR-2
LR-3
Market Name
Cod
Tilapia
Tilapia
Tilapia
Squid
Squid
Squid
Scallop
Pollock
Monkfish
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Skate
Mackerel
Mackerel
Tilapia
Blue Crab
Tilapia
Blue Crab
Snapper
Monkfish
Monkfish
Common Name
Cod, Pacific
Tilapia
Tilapia
Tilapia
Squid, Japanese Flying
Squid, Japanese Flying
Squid, Japanese Flying
Scallop, Sea
Pollock
Monkfish/Goosefish
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Skate, Winter
Mackerel, Atlantic
Mackerel, Atlantic
Tilapia
Blue Crab/Softshell
Tilapia
Blue Crab/Softshell
Snapper, Red
Monkfish/Gooseiish
Monkfish/Goosefish
Hg Concentration
(mg/kg)
0.029
0.011
0.015
0.011
0.13
0.071
0.02
0.021
0.019
0.023
0.019
0.017
0.036
0.009
0.038
0.009
0.032
0.11
0.11
Detected
Y
N
N
N
Y
Y
Y
N
Y
Y
Y
Y
N
Y
Y
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Reporting Limit
(mg/kg)
0.0099
0.0097
0.0099
0.0097
0.0084
0.0088
0.009
0.01
0.02
0.02
0.0092
0.0095
0.0091
0.0087
0.0086
0.0099
0.0096
0.0084
0.02
0.009
0.0093
0.0087
0.0088
0.02
0.02
-------
Table A-5. Mercury concentrations by composite sample number used in analysis (prior to averaging duplicates and
replicates) (continued)
Sample
Number
248.C
249.C
250.C
251.C
252.C
253.C
254.C
255.C
255.C
255.C
256.C
257.C
258.C
259.C
260.C
Sample
Group
248
249
250
251
252
253
254
255
255
255
256
257
257
259
260
Duplicate Of
257.C
Replicate ID
LR-1
LR-2
LR-3
Market Name
Skate
Whiting
Sea Bass
Snapper
Cod
Squid
Squid
Skate
Skate
Skate
Skate
Scallop
Scallop
Pollock
Blue Crab
Common Name
Skate, Winter
Whiting, Offshore
Bass, Black Sea
Snapper, Red
Cod, Atlantic
Squid, Longfin
Squid, Longfin
Skate, Winter
Skate, Winter
Skate, Winter
Skate, Winter
Scallop, Sea
Scallop, Sea
Pollock
Blue Crab/Hardshell
Hg Concentration
(mg/kg)
0.022
0.029
0.078
0.031
0.016
0.012
0.012
0.029
0.027
0.025
0.038
0.011
0.01
0.1
0.024
Detected
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Reporting Limit
(mg/kg)
0.0089
0.009
0.0099
0.0094
0.0096
0.0094
0.0086
0.0086
0.0088
0.009
0.0092
0.0088
0.0099
0.0091
0.0095
>
-------
APPENDIX B. SUMMARY OF DNA BARCODING ANALYSIS
B.I. INTRODUCTION
Traditionally biological specimens have been identified using morphological features.
For many species a trained technician can make routine identifications using morphological
features (taxonomic keys), but in many cases an experienced professional taxonomist is needed.
Even an experienced professional may not be able to identify a specimen in all cases. For
example, a specimen may be damaged or represent an immature stage of development, and thus
may lack key characters that allow it to be correctly placed taxonomically. Over the last 8 years,
"DNA barcoding" has emerged as a commonly employed tool that can help overcome some of
the issues associated with morphology based identifications. Barcoding uses a short genetic
sequence from a standard part of the genome (the total hereditary information of an organism,
encoded within its double-stranded DNA) in an attempt to accurately assign a specimen to a
given taxon, ideally, a species. Such an assignment can be made by examining a genomic region
(i.e., DNA sequence of base pairs made up of A's, C's, G's, and T's - see explanation below)
that exhibits a high degree of sequence conservation within a species, but appreciable divergence
compared to other species. An -650 base pair region of the mitochondrial gene cytochrome c
oxidase subunit 1 (coxl) has been found to commonly exhibit the requisite
conservation/divergence, and has been adopted as the standard "barcode" region for animals.
Barcoding is a valuable new tool in the taxonomist's toolbox that can supplement his or her
expert knowledge, and also provides a way for non-experts to make identifications (see
http://www.barcodeoflife.org/what-is-dna-barcoding).
Select samples from the FFM project had their COI barcode region sequenced as part of
the species identification process for the project. Barcoding was considered useful because the
sources of the fish specimens consisted of different commercial fishers with uncertain quality of
handling of the specimens and they were not from a scientific survey. Additionally, there was
concern that the handling of the fish specimens by the commercial fishers could result in damage
to the fish exterior that would make it difficult to identify to species level based on morphology
alone. In general it was believed that the use of DNA barcodes could help assure the quality of
specimen identification, and thus might impact the study results as a whole.
In order to use DNA barcodes to classify "unknown" specimens, the barcode sequences
from unknown specimens must be compared to "known" reference sequences. In 2005, the
Consortium for the Barcoding of Life (CBOL, see http://www.barcoding.si.edu), an international
consortium whose mission is to promote the exploration and development of DNA barcoding,
proposed a standard to be applied to sequences that are to be considered reference barcodes. The
standard was designed to apply to reference sequences deposited in the public domain, in
International Nucleotide Sequence Database Collaboration (INSDC) databases (GenBank, at the
US's National Center of Biotechnology Information [NCBI], the European Molecular Biology
Laboratory [EMBL], and the DNA Data Bank of Japan [DDBJ] [Wheeler et al., 2000; Benson et al.,
2000]). Background information about the initiative and the entire proposed standard can be
B-l
-------
found at http ://barcoding. si. edu/PDF/DWG data standards-Final .pdf. The details of the
standard are beyond the scope of this write-up; the point is that there is a significant and ongoing
international effort in this area and many species can now be reliably identified based on their
barcode sequence.
The University of Guelph in Ontario, Canada, under the leadership of Dr. Paul Hebert,
has established a publically available database and data management tool known as "BOLD
Systems" (Barcode of Life Database Systems; Ratnasingham and Hebert, 2007). Researchers the
world over have uploaded data into BOLD via the internet, making it a primary repository of
barcode data. BOLD also has an internet submission form that allows a sequence generated from
an "unknown" specimen to be submitted and compared to entries in BOLD (see
http://www.boldsystems.org/index.php/IDS OpenldEngine). Ideally the unknown sequence can
be assigned as belonging to a particular species based on its similarity to sequences in the
BOLD.
At the time of the FFM project, there were essentially three "levels" of the BOLD
database to which a test sequence could be compared. The most "stringent" level compared the
test specimen's sequence only to BOLD "reference" sequences, which were the sequences in
BOLD that met the CBOL reference barcode criteria described in the proposed standard
discussed above. If no "matches" of adequate similarity for species identification (roughly >98%
sequence identity) were found at that level, the search could be widened and allowed comparison
of the test sequence to reference and non-reference species records in BOLD. Non-reference
species records did not have all the associated elements required to be considered reference
sequences, but were still assigned to a particular species. If a test sequence was assigned to a
species based on its matching non-reference records, the result was considered more uncertain.
Finally, if no matches were found at either of the first two levels, a test sequence could be
compared to every record in BOLD, including those that were not classified down to the species
level. At this level of comparison, a test sequence would most likely be assigned as belonging to
a particular genus or family, but not a species. The other database queried for this project was
GenBank. Details about the specific process used during the FFM project to assign barcode-
based taxonomy will be presented later. The following paragraphs summarize, in basic terms,
the process used to obtain DNA barcode sequences, and provide additional conceptual
information about what they represent.
B.2. GENERATION OF DNA BARCODE SEQUENCES
The analytical process used to produce a barcode sequence starts by extracting and
purifying the DNA from a small aliquot (-20 mg) of tissue from a specimen, and then subjecting
the extracted DNA to an amplification process known as polymerase chain reaction (PCR). In
PCR, specific "primers" are used to target a particular genetic region, which is copied over and
over and over, resulting in geometric growth in the number of copies of the targeted region until
millions of copies of it are present in the final PCR product. The primers can target any region;
the primers used for the current project ('dgLCO1490', 'dgHCO2198'; Meyer, 2003) target the
B-2
-------
barcoding region of COL Following the PCR, an aliquot of the solution containing the amplified
target is subjected to a "sequencing reaction" in which fluorescent dyes are associated with the
nucleotides making up the target DNA, so that its sequence can be determined on an instrument
designed for that purpose.
DNA is a double-stranded molecule. Each strand is composed long chains of four
specific bases (nucleotides; adenine (A), guanine (G), cytosine (C), and thymine (T)) which
encode the hereditary information of an organism. For example, one strand of DNA, over a very
small stretch, might be composed of nucleotides in the order CTTAGGTGCA, and the ordered
letters representing the nucleotide bases are called the "sequence" of that stretch of DNA.
Importantly, the two strands making up a DNA molecule are complementary, due to the specific
pattern of hydrogen bonding that occurs between specific pairs of the four nucleotide bases; A
and T form stable hydrogen bonds with each other, and so do C and G. Based on this pattern of
hydrogen bonding, the complementary DNA strand for the example sequence above would be
GAATCCACGT. Figure B-l illustrates schematically how the example complementary
nucleotide sequences would line-up in the double-stranded stretch of DNA containing them.
CTTAGGTGCA
GAATCCACGT
strand
strand 2
Figure B-l. Schematic representation of complementary double-stranded
DNA.
The DNA barcode for an organism is the specific, ordered sequence of the nucleotide
bases (the A's, C's, G's, and T's) that are present in its COI barcode region. The sequencing
process was carried out on each of the complementary DNA strands so that the sequence of each
was determined independent of its complement. Due to their complementary nature, the two
sequences were then overlaid to confirm, with very high confidence, the identity of each position
in the resulting barcode sequence. The complementary nature of the strands also allows the
barcode sequence to be represented by a single sequence (strand). As previously described, once
a specimen's barcode sequence is determined, it can be compared to barcode sequences in public
databases (e.g., GenBank, BOLD), and based on its similarity to reference sequences in the
database(s), the specimen from which the sequence originated can be "assigned" to a species (in
most cases), or possibly to a higher taxonomic level (e.g., genus, family), depending on how
similar the specimen's barcode sequence is to the reference sequences.
B-3
-------
B.3. SAMPLING AND ANALYSIS OF FULTON FISH MARKET SAMPLES
The following paragraphs discuss some of the logistical and analytical aspects related to
the processing of the Fulton Market samples collected for DNA analysis. Tissue samples for
DNA analysis were collected at the Region 2 laboratory in Edison, New Jersey. Approximately
2 g of muscle tissue were taken from individual specimens designated for DNA analysis and also
from mixed specimen "super-composite" samples. Sampling tools were wiped clean with
ethanol between samples to prevent cross-contamination. Samples were stored frozen at -20°C
in "cryo-vials" until they were shipped to the DNA laboratory in coolers containing frozen ice
packs (results of a "test" shipment of two samples in April 2008 demonstrated that this form of
preservation for shipping allowed successful sequencing).
Five overnight shipments were sent from Edison to the National Exposure Research
Laboratory's Ecosystems Exposure Research Division (NERL/EERD) in Cincinnati between
June 12 and August 7, 2008; a total of 556 samples were shipped. Upon arrival in Cincinnati,
samples were visually inspected for signs of thawing and other problems, and verified against the
enclosed chain-of-custody forms. Samples were logged into a spreadsheet (Sample Log.xls)
located in the Fulton project's network directory in Cincinnati, and stored at -20°C until analysis.
No adverse sample conditions were noted in the sample log or on the custody sheets; these forms
were placed in the project lab notebook following sample receipt and inspection. At the
conclusion of the project, the Cincinnati sample log was verified as being correct and complete
with respect to the records from the shipping lab.
As described in Section 1.1, typically, three individual fish of the same species were
collected from a vendor to make up a composite sample for Hg and possibly PCB analyses.
When whole fish (rather than fillets) were collected, only one of the three was processed for
DNA sequence analysis (e.g., only sample 1.1 of samples 1.1, 1.2, 1.3 was sequenced), although
tissue samples from all three were collected and shipped to Cincinnati. Fillet samples were
handled differently. All three samples collected for the Hg analysis were processed for DNA
sequencing as visual identification of a species from a fillet sample is often difficult. The final
type of sample subjected to DNA analysis was the single aliquots taken from the mixed-tissue
super-composite samples. Employing this analysis plan yielded a list of 282 samples for DNA
analysis. Six additional samples were added to the analysis list later. The added samples were
three pairs of "x.2" and "x.3" samples, which were sequenced so the results could be compared
with their corresponding "x.l" results, which had some uncertainty or problems associated with
them. The total number of samples to be sequenced was therefore 288, 284 of which were
successfully sequenced.
One of the four unreported samples, 222.1, twice produced sequences that appeared to be
of bacterial origin, so no further attempts were made to analyze that sample. Attempts to
generate results for the other three unreported samples (141.1, 176.1. and 233.1) were never
successful, despite repeated attempts at DNA extraction, PCR, and sequencing.
B-4
-------
B.4. QUALITY ASSURANCE AND QUALITY CONTROL
PCR and sequencing samples are typically processed in 96-well plates (8 rows * 12
columns), and a plate was considered an analytical "batch" regardless of the number of samples
loaded on the plate. During this project, negative controls (no DNA template), positive controls
(known specimen), and replicate samples were included for QC. Though not considered a
critical omission, negative and/or positive control samples were not included in a few batches;
see the explanation below.
Negative controls serve to indicate possible contamination. Reasons why a negative
control might produce a sequence include: (1) the sample was contaminated with environmental
DNA that carried over from harvest, fish market, or processing; (2) the sample was contaminated
with human, fungal, or bacterial DNA at some step between harvest and extraction; (3) the DNA
was contaminated with volatilized PCR products from previous PCR runs in the same lab; or (4)
samples were mislabeled or mis-ordered. For this project, if a negative control produced a
sequence, the sequencing reaction product was re-sequenced to see if the sequence remained or
not (i.e., was an artifact of the sequencing process). If the second sequencing run confirmed the
presence of a sequence in the negative control well, then the entire batch (plate) of samples
associated with that control was rejected and reprocessed. There was only one case where a
plate was rejected due to negative control failure as described.
Regarding contamination, it should also be noted that human DNA sequence was found
twice, however, repeating the PCR amplification from the original DNA extract from the two
samples in question yielded sequences that were more appropriately assigned to fish species.
Based on these results, it is considered most likely that human contamination somehow occurred
during the original PCR or sequencing reactions for these samples
Though the QA plan for the project indicated the use of positive controls and replicates in
every sample batch, they were only used sporadically. Positive controls were included in the
PCR process, but were omitted from sequencing. Because all of the samples included in this
project were identified morphologically, the lack of positive controls is not considered critical
since all samples' resulting barcode sequences were used to confirm morphology-based
identifications. Details about replicate samples that were analyzed can be found in Table B-l,
which summarizes batch/plate analyses and the associated QC samples and results.
The minimum acceptable length of each of the two complementary DNA sequences
generated for a given specimen was 500 bases, with a minimum overlap of the complementary
sequences of 400 bases. (Typically, the total length of the final sequence for any specimen, once
the two overlapping sequences were combined, was approximately 665 bases.) There were a few
cases where a reported result was based on sequences that did not meet the length criteria. In
two cases, one of the individual DNA sequences was <500 bases, and in the other the overlap
was <400 bases. In both cases, the overall quality of the individual reads and resulting final
sequences and the agreement of the final sequences with reference sequences lead to the decision
not to reanalyze the samples. In a third case, a result based on only one of the complementary
DNA sequences from a specimen was reported, due to difficulties experienced in sequencing the
B-5
-------
second strand. This case represented a "confirmatory" re-analysis (discussed below), and the
single sequence agreed with the morphologically-determined species assignment, so further
analysis was deemed unnecessary. In the deliverable to Region 2 containing the record for this
sample (44.1), the comment field noted that the result was based on a single DNA strand.
Additional QA/QC procedures related to the DNA analysis are described in the Quality
Assurance Project Plan (QAPP) for the project.
-------
Table B-l. DNA analysis, reporting, and QC summary
Batch/Plate
NY1
NY2
NY3
NY4.1
NY5
NY_Fillets
#of
Samples
30a
64
42
58
42
46
# Submitted
for
Sequencing
10b
64C
42
58
42
46
#of
Samples
Reported
3
53
37
51
38
40
#of
Negative
Controls
la
1
1
1
1
2
Did Negative Control Sequence?
No
No
No
No
Not submitted for sequencing; no band
on agarose gel
Negative control sequenced in forward
direction on first run, however did not
sequence on reanalysis, so OK
#of
Replicates
la
2C
1
1
1
0
Replicate % Identity
100%
100%, 100%
100%
100%
Not submitted for sequencing;
appropriate band on agarose gel
N/A
Extraction and PCR repeats
NYrepeats#5d
Nyrepeats#6e
Nyrepeats#7 and
NYGrad#lf
Nyrepeats#8
TD1290Lg
15
16
6/7
21
2
15
16
6/7
21
2
0
0
0/2
12
2
1
1
1/1
1
0
Negative control sequenced both
directions - all results discarded
No
Control not submitted for sequencing;
very faint band on agarose gel; however
not considered critical given that only 2
results reported and both match morph.
ID.
Negative control sequenced in forward
direction on first attempt, however did
not sequence on reanalysis, so OK
N/A
0
0
0/0
0
2
N/A
N/A
N/A/N/A
N/A
100%, 100%
PCR only repeats
NY1.2
NY1. 3 (50 degrees)
NY1.4(Deg. Folmers)
30a
30a
30a
llb
2b
15C
6
0
11
la
la
la
No
No
No
la
la
T
100%
100%
100%, 100%
td
-------
Table B-l. DNA analysis, reporting, and QC summary (continued)
Batch/Plate
NY2and3_repeats
NY1.4and2_repeats
NYrepeats dg
NYChi+repeats
NYrepeats#5d
NYrepeats#5cd
Samples
8
15
20
6
9
24
# Submitted
for
Sequencing
6
15
20
6
9
24
#of
Samples
Reported
5
10
10
1
0
3
#of
Negative
Controls
0
1
1
1
1
1
Did Negative Control Sequence?
N/A
No
No
Negative control sequenced in reverse
direction on first attempt, however did
not sequence on reanalysis, so OK
Negative control sequenced both
directions - all results discarded
No
Replicates
0
0
1
1
0
0
Replicate % Identity
N/A
N/A
Replicate failed to produce
sequence, so not evaluated
100%
N/A
N/A
aNYl, NY1.2, NY1.3, and NY1.4 were all the same sample extracts amplified with different PCR primers and/or with different PCR conditions in an effort to find the best PCR
td conditions. Negative controls and replicates in those plates represent the same negative control and replicate extracts.
oo bA total of 19 single samples plus two each of 13.1 and 50.1 (from different PCR plates) were submitted for sequencing from plates NY1,NY1.2, andNYl.3. These 23 total
samples were sequenced together. Of these 23 samples, nine were reported. Only one of the two 50.1 samples produced good sequence. Both 13.1 samples produced good
sequence, so these results were used as a substitute for the failed replicate of 53.1, which had only one of its two replicate extractions (the second) produce good sequence. Though
used as a replicate for QC purposes in this batch (reported as the "task_1290b" deliverable), the 13.1 results from this batch were not reported. Another aliquot of sample 13.1 from
plate 1.4 was also analyzed, and the final results from that analysis were the ones reported (in the "task_1290c" deliverable).
°NY2 and NY1.4 were submitted for sequencing together, with the negative control and replicate from plate NY2 (sample 150.1) serving as primary QC samples. The first
extraction aliquot of sample 53.1 was submitted again as part of plate NY1.4, but is not counted in the # of Samples Reported Column since its result was previously reported,
however it is counted as an additional replicate for plates NY2 andNY1.4. It shared 100% identity with the previously reported 53.1.
dPlate NYrepeats#5 was a combination of nine samples subject only to re-amplification by PCR and another 15 samples that were both re-extracted and re-amplified. The negative
control on this plate sequenced in both directions. All 24 extracts and the negative control were subjected to PCR again, and this time the negative control showed no sign of
contamination. This "clean" plate was submitted for sequencing and was named NYrepeats#5c.
eAll 16 samples on NYrepeats#6 were confirmatory analyses of samples previously reported.
fThese were two PCR repeat plates sequenced together.
gThe analysis of these two samples (129.1 and 164.1) was unique because previous rounds of PCR had resulted in multiple bands (products) appearing on the post-PCR agarose gel.
In task!290L, following PCR and gel electrophoresis, the bands of the appropriate size (i.e., the assumed target) were excised from the gel and purified. This purified product was
then sequenced and was confirmed to represent the target, and the results reported.
hAll "repeat" plates may contain combinations of samples which were previously unreported and samples that were previously reported to Region 2 that were being reanalyzed for
some reason (such as being identified as a different species genetically compared to the morphologically assigned species designation).
-------
B.5. ASSIGNING DNA-BASED TAXONOMY
The following paragraphs describe the process used on this project to make DNA
barcode-based taxonomic assignments. In the simplest terms, a specimen was assigned to a
particular species when its barcode sequence strongly matched (approximately >98% identical)
the sequence of one or more reference sequences of a single species in GenBank and/or BOLD.
However, there were cases where this did not occur. Matches of a lower percentage identity
typically resulted in assigning the "unknown" specimen to a genus or family rather than to a
species. There are other reasons a specimen may not have been assigned to a particular species,
which are discussed below.
Following the generation of sequence data for any given batch of samples, a simple text
file was output from the sequence editing program (Sequencher, v4.8, www.genecodes.com)
containing a sample identifier and the associated barcode sequence for each sample in the batch.
Using a web browser, the file was uploaded and compared to the sequences in GenBank using
the "megablast" algorithm (Zheng et al., 2000; a particular version of the Basic Local Alignment
Search Tool, BLAST, http://www.ncbi.nlm.nih.gov/blast). The BLAST results returned via the
browser report which GenBank sequences most closely match the "query" sequences in the
uploaded file. A tab-delimited text file summarizing the results, called a "hit table," was
downloaded from the BLAST results page and saved, and then a custom Perl script (computer
program; Wall et al., 2000) was used to parse the results file and to retrieve taxonomic
information associated with each of the top ten BLAST "hits" for any given input sequence. The
taxonomy results were culled from the taxonomy database associated with GenBank
(http://www.ncbi.nlm.nih.gov/taxonomy) and augmented with information from the Integrated
Taxonomic Information System (ITIS, originally referred to as the Interagency Taxonomic
Information System; www.itis.gov). The script also automatically went back to GenBank and
examined the full GenBank records of the "hit" sequences to see if they contained the
"BARCODE" designation and to see if they were cross-referenced to records in BOLD. In
theory, having one or both of these features would provide additional confidence in the veracity
of the assigned taxonomy of the "hit" sequence.
Following the compilation and summarization of the BLAST-based results, the next step
was to go to the BOLD "species identification" page,
http://www.barcodinglife.org/views/idrequest.php. Here, each individual query sequence from
the analytical batch was cut and pasted into the webpage form and submitted for species
identification using the proprietary BOLD algorithm. As discussed previously, the query
sequence was first compared to BOLD reference sequences (the default choice in the web form),
but if no species match was returned, the query sequence was compared to the non-reference
species database. If there were still no species-level results, then the "entire database" option
was used for the query.
Once all the results from both GenBank and BOLD were gathered, a weight-of-evidence
approach was used to make a DNA-based taxonomic assignment. When the BOLD and
GenBank results agreed (i.e., BOLD assigned a species-level identification that was the same
B-9
-------
species associated with the best GenBank results), a sample could be assigned to that species
with relatively high confidence. The confidence in such a result was greater if the BOLD
"match" was to one or more reference sequences and/or the GenBank/BLAST "hits" were to
official "BARCODE" sequences.
In some cases however, there was not such a clear correspondence between the BOLD
and GenBank results. For example, a BOLD species identification might be relied on
exclusively when there were no corresponding BLAST hits (i.e., COI barcode sequence data for
the species in question was present in BOLD, but not available in GenBank). The opposite
might also be true, so that a taxonomic assignment for a particular sample might be based only
on comparison to sequences found in GenBank. As previously stated, a species-level assignment
sometimes was not made because of lower than ideal sequence similarity results, so a genus-
level or family-level assignment was made. There were also cases where GenBank and BOLD
both provided apparent species level matches, but they did not agree. Finally, there are some
groups (e.g., tuna, tilapia, most snappers) where there doesn't appear to be enough divergence in
COI barcode region to reliably make a species assignment. In the end, 195 of the 284
successfully sequenced samples were reported to species level. Regardless of the taxonomic
level assigned, a column in the deliverables to Region 2 was used to indicate the primary
database (BOLD reference, BOLD non-reference, or GenBank) that was used for taxon
assignment; GenBank and BOLD were both referenced when the supporting evidence came from
both databases and appeared to be independent.
B.6. COMPARISON WITH MORPHOLOGICALLY DERIVED TAXONOMY
RESULTS
Samples were initially sequenced and "DNA identified" blindly, and the results submitted
to Region 2 in an agreed upon spreadsheet format. (It should be noted that the DNA sequences
themselves were not delivered to Region 2. Rather, Region 2 received a single-line record
summarizing the results of the "unknown" sequence's comparison to GenBank and BOLD, and
its inferred taxonomic classification.) Upon receipt of deliverables, Region 2 compared the
sequence-based results to their own morphologically-based taxonomic assignments. In cases
where there was disagreement, NERL/EERD was notified, and the samples were subjected to
reanalysis to confirm or refute the earlier DNA results. In total, 27 samples were reanalyzed due
to such disagreements (or in a few cases for other reasons). In five of these cases, the "DNA ID"
was changed after reanalysis and agreed with the morphologically-based result. For two of the
modified DNA results, it appeared likely that the initial error was due to two samples' positions
being switched on the 96-well processing plate during the original analysis.
As stated above, difficulties experienced when attempting to sequence one of the two
DNA strands for another sample (number 44.1) resulted in using a single strand's sequence to
assign the sample's taxon, which agreed with the morphologically-based assignment. There is
no clear explanation for the disagreement between the DNA-based results and morphologically-
B-10
-------
based results for the other two samples. Most likely, cross-contamination occurred somewhere
in the sampling or analytical process, which lead to erroneous sequencing results.
Of the remaining 22 reanalyzed samples, one clearly had incorrect DNA results; the
photograph for sample 232 was clearly a scallop, however the DNA results twice came up as a
cod. It is again assumed there must have been a handling error in the field or lab that caused
cross-contamination, perhaps the wrong tissue was placed in the sample vial or the vial was
mislabeled. Another case involved a sample representing fish that typically resolved to species
level resolving only to genus level based on its initial DNA analysis; the reanalysis did not
change this. (Region 2 still assigned this sample to its presumed species based on its
morphology.) Another five samples (four fillet samples and one mussel) were re-sequenced due
to initial DNA results that were confounded or were of poor quality (including one case
evidencing human contamination). The five samples reconciled with their market names based
on the reanalysis.
Of the remaining 15 reanalyzed samples, 14 had their final species changed, including
one changed from a striped bass to a striped bass hybrid. The fifteenth, though not officially
changed, was flagged as potentially being another species (see note b, Table 1). Not only
reanalyzed samples were changed however, as the entire process of DNA analysis, reanalysis,
and reexamination of photos by Region 2's ichthyologist, Moses Chang, resulted in a total of 57
samples (not considering the three "sub-samples" making up the composite samples),
representing 15 species, having their final identification changed (or in some cases where there
was initial uncertainty, confirmed). Table B-2 summarizes the changes that were made.
It is important to note that scientific names were changed, but not the market names
(though FDA vernacular names often differed). As people purchase fish based on market names,
and considering the analysis presented in Section 3.1 of this report, it appears likely that market
customers do not face a great risk of obtaining improperly labeled seafood containing higher
levels of mercury, however the small sample size for any given species in this study makes it
hard to make that assertion with any confidence. Taking that into account and given that there
were a significant number of species identifications changed based on the DNA, future studies of
this type should still consider the use of DNA barcoding to allow a better examination of the
potential for species specific risks. Aside from that consideration, the FDA, in association with
the Consortium for the Barcode of Life (CBOL) and its FISH-BOL project (www.fishbol.org)
developed a laboratory protocol for DNA barcoding offish (LIB 4420: DNA-Barcoding for the
Species Identification of Fish, July, 2008, updated July 2009 (Yancy, 2008). FDA has since
developed an official Standard Operating Procedure (SOP) for DNA barcoding offish to replace
LIB4420, "Single Laboratory Validated Method for DNA-Barcoding for the Species
Identification of Fish for FDA Regulatory Compliance"
(http://www.fda.gov/Food/ScienceResearch/LaboratoryMethods/ucm237391 .htm). Along with
the development of the SOP, the "FDA Reference Standard Sequence Library for Seafood
Identification" has been established (http://www.fda.gov/Food/FoodSafety/Product-
SpecificInformation/Seafood/DNAspeciation/ucm238880.htm). As the use of DNA barcodes to
B-ll
-------
ensure proper identification of seafood is now part of the regulatory landscape, its use in the
current study aligned well with the general trends in this arena.
B-12
-------
Table B-2. Influenced species identifications in final project database
Sample
ID#s
13
78
229
230
231
28
52
68
89
131
151
167
217
219
220
238
248
255
256
Species Name in Project
Database
Original
Loligo
pealii
Loligo
pealii
Raia
binoculata
(•)
Final
Illex spp.
Todarodes
pacificus
Leucoraja
ocellata
Common Name in Project
Database
Original
Squid
Squid
Skate
Final
Squid
Squid,
Japanese
Flying
Skate, Winter
Acceptable FDA Market Names
Original
Squid or Calamari
(vernacular; Winter Squid/
Boston Squid/ Longfm
Inshore Squid)
Squid or Calamari
(vernacular; Winter Squid/
Boston Squid/ Longfm
Inshore Squid)
Skate
Final
Squid or Calamari
(vernacular; various
depending on species;
none in common with L.
pealii)
NA
Skate
td
I
U)
-------
Table B-2. Influenced species identifications in final project database (continued)
Sample
m-Wc.
tfS
33
43
64
173
193
40
199
48
Species Name in Project
Database
Original
Protothaca
Protothaca
Protothaca
Syacium
micrurum
Final
Mercenaria
mercenaria
Mercenaria
mercenaria
Mercenaria
mercenaria
Pseudopleur
o-nectes
americanus
Common Name in Project
Database
Original
Clam, Littleneck
Clam, Cherry
Stone
Clam, Top Neck
Flounder/Sole,
Channel
Final
Quahog,
Northern/Little
neck Clam
Quahog,
Northern/
Cherrystone
Clam
Quahog,
Northern/Top
Neck Clam
Flounder,
Blackback
Acceptable FDA Market Names
Original
P. thaca - Clam, Hardshell
or Quahog
P. staminea - Clam,
Littleneck (vernacular;
Steamer/ Native
Littleneck)
P. tenerrima -
Clam, Littleneck
P. thaca - Clam, Hardshell
or Quahog
P. staminea -Clam,
Littleneck (vernacular;
Steamer/ Native
Littleneck)
P. tenerrima -
Clam, Littleneck
P. thaca - Clam, Hardshell
or Quahog
P. staminea -Clam,
Littleneck (vernacular;
Steamer/ Native
Littleneck)
P. tenerrima -
Clam, Littleneck
NA
Final
Clam or Quahog
(vernacular; Hardshell/
Littleneck)
Clam or Quahog
(vernacular; Hardshell/
Littleneck)
Clam or Quahog
(vernacular; Hardshell/
Littleneck)
Flounder or Sole
(vernacular; Winter
Flounder/ Lemon Sole/
Georges Bank Flounder)
td
-------
Table B-2. Influenced species identifications in final project database (continued)
Sample
ID#s
138
162
51
67
83
109
148
186
221
259
57
84
100
106
156
85
107
149
Species Name in Project
Database
Original
Paralichthy
s dentatus
7
Pollachius
pollachius
Penaeus
monodon
Ocyurusl
chrysurusl
7
Final
Pseudopleur
o-nectes
americanus
Pseudopleur
o-nectes
americanus
Pollachius
virens
Litopenaeus
vannamei
Ocyurus
chrysurus
Rhomboplit
es
aurorubens
Common Name in Project
Database
Original
Flounder/Sole,
Summer
Flounder/Sole,
Canadian BB
Pollock, Atlantic
Shrimp, Black
Tiger
Snapper,
Yellowtail
Snapper, B-Liner
Final
Flounder,
Blackback
Flounder,
Blackback
Pollock
Shrimp, White
Snapper,
Yellowtail
Snapper,
Vermilion
Acceptable FDA Market Names
Original
Flounder or Fluke
(vernacular:
Plaice/Northern Fluke)
NA
Pollock (vernacular;
Lythe/ Saithe/ Dover
Hake/ Grass Whiting/
Greenfish/ Margate Hake)
Shrimp (vernacular;
Jumbo Tiger Prawn/
Black Tiger Shrimp)
Snapper (vernacular: Palu-
i'usama)
NA
Final
Flounder or Sole
(vernacular; Winter
Flounder/ Lemon Sole/
Georges Bank Flounder)
Flounder or Sole
(vernacular; Winter
Flounder/ Lemon Sole/
Georges Bank Flounder)
Pollock (vernacular;
Saithe/ Coalfish/ Coley/
Green Cod/ Boston
Bluefish)
Shrimp
Snapper (vernacular: Palu-
i'usama)
Snapper (vernacular:
Beeliner/ Clubhead
Snapper/ Night Snapper)
td
-------
Table B-2. Influenced species identifications in final project database (continued)
Sample
ID#s
98
195
108
166
114
142
174
180
249
222
223
224
Species Name in Project
Database
Original
Morone
saxatilis
Morone (?)
saxatilis (?)
Ictalurus
punctatus
Ictalurus
punctatus
Ictalurus
punctatus
Crassostrea
virginica (?)
Merluccius
bilinearis
Gadus
morhua
Final
Morone
chrysops x
saxatilis
Morone
chrysops x
saxatilis
Ameiurus
catus
Ameiurus
catus
Ictalurus
furcatus
Crassostrea
gigas
Merluccius
albidus
Gadus
macrocepha
lus
Common Name in Project
Database
Original
Striped Bass
Striped Bass,
Hybrid
Catfish
Catfish, Channel
Catfish
Oyster
Whiting
Cod
Final
Bass, Hybrid
Striped
Bass, Hybrid
Striped
Catfish, White
Catfish, White
Catfish, Blue
Oyster, Pacific
Whiting,
Offshore
Cod, Pacific
Acceptable FDA Market Names
Original
Bass (vernacular;
Rockfish/ Striper/
Linesides)
Bass
Catfish (vernacular;
Spotted Cat/ White Cat/
Lake Catfish)
Catfish (vernacular;
Spotted Cat/ White Cat/
Lake Catfish)
Catfish (vernacular;
Spotted Cat/ White Cat/
Lake Catfish)
Oyster (vernacular;
Bluepoint/ American
Oyster)
Whiting (vernacular;
American Hake/
Silverfish/ Stockfish/
Winter Trout/ Frostfish)
Cod (vernacular; Rock
Cod/ Codling/ Scrod Cod)
Final
Bass (vernacular; White
and Striped Bass Hybrid)
Bass (vernacular; White
and Striped Bass Hybrid)
Catfish (vernacular; White
Cat/ Channel Cat)
Catfish (vernacular; White
Cat/ Channel Cat)
Catfish (vernacular;
Bullhead/ Chuckleheaded
Cat/ Mississippi Cat)
Oyster (vernacular;
Japanese Oyster/ Pacific
Giant Oyster)
NA
Cod or Alaska Cod
(vernacular; Alaska Cod/
Grey Cod/ True Cod/
Treska)
td
-------
B.7. RESULT QUALIFIER SCORE
At the conclusion of the DNA sequencing effort, a "qualifier score" was created for each
reported sample record to reflect some estimate of the confidence associated with the result. The
score also reflected the result's relative taxonomic position, as a given result would start with a
qualifier score of 0, 0.5, or 1 depending on whether or not its DNA-based taxonomic level was
family, genus, or species, respectively. Without going into detail about the empirically-derived
scoring criteria, it can be said that a species-level assignment would generally be expected to be
higher scored than a genus-level assignment, and a genus-level assignment higher than a family-
level assignment. Various weight-of-evidence criteria such as mentioned above were used to
adjust the score. For instance, sequences matching the same taxa in both GenBank and BOLD
sequences would be higher scored, with additional weight given when matches were to GenBank
"BARCODE" sequences and/or BOLD "reference" sequences. Higher percentage shared
identity with database sequences also increased the qualifier score. The qualifier score for any
sample ended up somewhere between 0 and 5, where a lower score indicated the sample was
more likely assigned to a higher taxonomic level (e.g., family) with less weight of evidence, and
a higher score indicating that the sample was probably assigned to a species, with reasonably
strong evidence supporting its assignment.
As mentioned above, the qualifier score was derived empirically, and was not intended to
be a robust measure of result quality. Rather, it was an attempt to give end-users, particularly
those unfamiliar with DNA barcoding, some idea about which results were the best and which
might be improved upon. There might not be any easy way to "improve" a given result at the
current time however, given that a number of factors are out of the control of a person using a
barcode sequence to identify an "unknown" specimen. One attempting identification is always
subject to what is available in the reference databases and the quality of the database contents
(e.g., Is the species they are trying to identify represented in the databases? Are the sequences in
the database correct? Were the specimens the sequences came from identified correctly in the
first place?). Even if reference COI barcode sequences are available, this gene region alone does
not necessarily resolve all animal species, as was seen in the present study. A detailed
description of the calculation of the qualifier score is beyond the scope of this report, however it
can be stated that generally more weight was given to results based on sequences that matched
GenBank "BARCODE" and/or BOLD "reference" sequences, and by giving more weight to
taxonomic assignments that were based on matching independent sequences in both databases.
Table B-3 displays the count and average qualifier score for different taxonomic levels
based on the databases used for their assignment. It is worth pointing out that at the species
level, GenBank/BOLD non-reference-based results have a higher average qualifier score than
GenBank/BOLD reference-based results. This is surprising, as it implies that many BOLD "non-
reference" sequences that were matched had more supporting evidence in GenBank than BOLD
"reference" sequences that were matched. It is assumed this is the result of chance, however a
detailed review to confirm this assumption was not performed.
B-17
-------
B.8. ADDITIONAL INFORMATION
More details about any aspect of the DNA barcoding analysis, including specific
laboratory procedures, sequence data files, summarized BLAST/BOLD results, deliverable files
submitted to Region 2, qualifier score calculation, etc. can be obtained by contacting John
Martinson (martinson.john@epa.gov).
Table B-3. Distribution of taxa level assignments based on taxon assignment
source database(s) and average result qualifier score
Assigned taxonomic level
family
Taxa Assignment Source
BOLD non-reference
Genbank/BOLD reference
Data
Average of total score
Count of Assigned_taxonomic_level
Average of total score
Count of Assigned_taxonomic_level
family Average of total score
family Count of Assigned taxonomic level
genus
BOLD non-reference
Genbank
Genbank/BOLD non-reference
Genbank/BOLD reference
Average of total score
Count of Assigned_taxonomic_level
Average of total score
Count of Assigned_taxonomic_level
Average of total score
Count of Assigned_taxonomic_level
Average of total score
Count of Assigned_taxonomic_level
genus Average of total score
genus Count of Assigned taxonomic level
species
BOLD non-reference
BOLD reference
Genbank
Genbank/BOLD non-reference
Genbank/BOLD reference
Average of total score
Count of Assigned_taxonomic_level
Average of total score
Count of Assigned_taxonomic_level
Average of total score
Count of Assigned taxonomic level
Average of total score
Count of Assigned_taxonomic_level
Average of total score
Count of Assigned_taxonomic_level
species Average of total score
species Count of Assigned taxonomic level
Total Average of total score
Total Count of Assigned taxonomic level
Total
0.5
10
1.1
10
0.8
20
1.0
10
0.8
1
0.8
29
2.7
29
1.65
69
1.5
23
2.5
43
2.5
8
3.8
61
3.3
60
3.0
195
2.5
284
B.9. ACKNOWLEDGEMENTS
The efforts of Tamara Goyke (NERL/EERD), Stephen Morris (independent student
contractor), and Carrie Drake (Dynamac Corp.) are greatly appreciated. They all played key
roles and contributed greatly to the success of the effort to generate the DNA sequence data for
this project. Suzanne Christ, Erik Pilgrim, John Darling, and Eric Waits are thanked for their
helpful review comments.
B-18
-------
B.10. REFERENCES
Benson, DA; Karsch-Mizrachi, I; Lipman, DJ; Ostell, J; Rapp, BA; Wheeler, DL (2000).
GenBank. Nucleic Acids Res 28(1): 15-18.
Meyer, CP (2003). Molecular systematics of cowries (Gastropoda: Cypraeidae) and 538
diversification patterns in the tropics. Biol J Linnean Soc 539(79):401-459.
Ratnasingham, S; Hebert, PDN (2007). BOLD: The Barcode of Life Data System (www.
barcodinglife.org). Mol Ecol Notes, doi: 10.1111/j.l471-8286.2006.01678.x.
Wall, L; Christiansen, T; Orwant, J (2000). Programming Perl, 3rd ed, Sebatopol, CA: O'Reilly
Media.
Wheeler, DL; Chapper, C; Lash, AE; Leipe, DD; Madden, TL; Schuler, GD; Tatusova, TA;
Rapp, BA (2000). Database resources of the National Center for Biotechnology
Information. Nucleic Acids Res 28(1): 10-14.
Yancy, HF; Fry, FS; Randolph, SC; Deeds, J; Ivanova, NV; Grainger, CM; Hanner, R; Weigt,
LA; Driskell, A; Hunt, J; Ormos, A; Hebert, PDN (2008/2009). A Protocol for Validation
of DNA-Barcoding for the Species Identification of Fish for FDA Regulatory
Compliance, Lab Inform Bull 4420(24), Division of Field Science, Office of Regulatory
Affairs, U.S. Food and Drug Administration.
Zhang, Zheng; Schwartz, S; Wagner, L; Miller, W (2000). A greedy algorithm for aligning DNA
sequences. J ComputBiol 7(1-2):203-214.
B-19
-------
APPENDIX C. PCB ANALYSIS
A subset of the CM composite samples was also analyzed for the presence of
polychlorinated biphenyls (PCBs). The PCB analysis was limited to a maximum sample number
of 50 due to cost constraints. Given the limited sample size, five species (likely to be high in
PCBs) were initially selected for analysis. Two samples (one scallop, one swordfish) selected
only for mercury analysis were also analyzed for PCBs. Additionally, Spanish mackerel is listed
separately from mackerel. Consequently, the 49 samples selected for the analysis, as shown in
Table C-l represent a total of eight different species instead of five. The analysis tested for 124
different PCB congeners, as shown in Table C-l, and the analysis of each congener had an
associated limit of detection. The resulting PCB concentrations from each congener were then
summed to represent an estimate of the total PCB in the sample. Three of the samples were
analyzed in triplicate as a further test of laboratory analysis precision. Twelve PCB congeners
have dioxin-like toxic equivalency factors (TEFs). Not all of the 12 dioxin-like PCB congeners
were analyzed, and, consequently, no attempt was made to estimate a dioxin TEF from the PCB
sampling analysis.
As shown in Table C-2, many of the congeners resulted in analytical results below the
limit of detection. In this case, the limit of detection represents a sample-specific reporting limit.
Sample-specific limits take into account both the preparation and analysis effects as well as
specifics related to the specific sample matrix such as percent dry weight, grams of sample
weighed out, and exact solvent volumes used. These limits are intended to indicate the level at
which the method can reliably quantify positive results, where reliability is measured in terms of
the method precision and accuracy limits.
When a sample result is below this limit, an assumption must be made about the
approximate magnitude before the different congeners are summed to estimate the total PCB
concentration. However, because so many of the congeners had non-detected levels of PCBs,
adding even one-half the detection limit could severely overestimate the total PCB concentration.
In this analysis, species which had no detectable levels in any congeners were not quantitatively
analyzed; for species with detected levels in at least one congener for all samples (catfish), the
assumption was made that the levels in the other congeners were zero. This assumption tended
to underestimate the total PCB concentrations but likely represents the method with the smallest
bias. In the three samples examined in triplicate, all congeners returned non-detected levels in
all three trials.
Table C-3 shows the total PCB concentrations in the eight catfish samples. The
concentrations are presented in two ways: (1) as mass of PCB per mass offish tissue, and (2) as
mass per lipid in the tissue. The measured lipid fraction is used to convert the PCB mass per
total sample weight into a mass per lipid value. The current FDA tolerance level for PCBs in
fish is 2 ppm fish tissue (U.S. FDA, 2001) or 2,000 |J,g/kg. In the CM catfish samples, none of
C-l
-------
Table C-l. PCB congeners evaluated in analysis
PCB Number
PCB1
PCB 4
PCB 5
PCB 6
PCB 7
PCB 8
PCB 9
PCB 10
PCB 12
PCB 14
PCB 15
PCB 16
PCB 17
PCB 18
PCB 19
PCB 22
PCB 24
PCB 25
PCB 26
PCB 28
PCB 29
PCB 32
PCB 34
PCB 35
PCB 40
PCB 41
PCB 44
PCB 45
PCB 46
PCB 48
PCB 49
PCB 51
PCB 52
PCB 54
PCB 64
PCB 67
PCB 69
PCB 70
PCB 71
PCB 73
PCB 74
PCB 75
PCB 77
PCB 81
PCB 82
PCB 87
PCB 95
PCB 97
PCB 99
PCB 100
PCB 105
PCB 110
PCB 115
PCB 117
PCB 122
PCB 124
PCB 128
PCB 130
PCB 132
PCB 134
PCB 135
PCB 136
PCB 137
PCB 138
PCB 141
PCB 144
PCB 147
PCB 149
PCB 151
PCB 153
PCB 156
PCB 157
PCB 158
PCB 163
PCB 164
PCB 165
PCB 167
PCB 170
PCB 172
PCB 173
PCB 174
PCB 175
PCB 176
PCB 177
PCB 179
PCB 180
PCB 183
PCB 185
PCB 187
PCB 189
PCB 190
PCB 191
PCB 193
PCB 194
PCB 196
PCB 197
PCB 199
PCB 200
PCB 202
PCB 203
PCB 205
PCB 206
PCB 208
PCB 209
PCB 13 /PCB 27
PCB 31 /PCB 53
PCB 20 / PCB 33
PCB 47 /PCB 104
PCB 42 / PCB 59
PCB 37 /PCB 103
PCB 63 / PCB 93
PCB 66 /PCB 91
PCB 56 /PCB 84 /PCB 92
PCB 60 /PCB 90 /PCB 101
PCB 83 /PCB 119
PCB 85 /PCB 154
PCB 109 /PCB 123
PCB 118 /PCB 131
PCB 114 /PCB 146
PCB 129 /PCB 178
PCB 171 /PCB 201
PCB 195 /PCB 207
C-2
-------
Table C-2. Samples measured for PCBs and the number of congeners with
detected and non-detected levels of PCBs
Market Name and
Sample Number
Catfish, Sample 1
Catfish, Sample 2
Catfish, Sample 3
Catfish, Sample 4
Catfish, Sample 5
Catfish, Sample 6
Catfish, Sample 7
Catfish, Sample 8
Crab, Blue, Sample 1
Crab, Blue, Sample 2
Crab, Blue, Sample 3
Crab, Blue, Sample 4
Crab, Blue, Sample 5
Crab, Blue, Sample 6
Crab, Blue, Sample 7
Crab, Blue, Sample 8
Crab, Blue, Sample 9
Crab, Blue, Sample 10
Mackerel, Sample 1
Mackerel, Sample 2
Mackerel, Sample 3
Mackerel, Sample 4
Mackerel, Sample 5
Mackerel, Sample 6
Mackerel, Sample 7
Mackerel, Spanish,
Sample 1
Mackerel, Spanish,
Sample 2
Number of
Non-Detects
76
97
119
117
117
112
119
119
121
122
122
122
122
122
122
122
122
122
122
122
122
122
122
122
122
122
122
Number of
Detects
46
25
3
5
5
10
3
3
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Market Name and Sample
Number
Salmon, Atlantic, Sample 1
Salmon, Atlantic, Sample 2
Salmon, Atlantic, Sample 3
Salmon, Atlantic, Sample 4
Salmon, Atlantic, Sample 5
Salmon, Atlantic, Sample 6
Salmon, Atlantic, Sample 7
Salmon, Atlantic, Sample 8
Salmon, Atlantic, Sample 9
Salmon, Atlantic, Sample 10
Scallop, Sample 1
Swordfish, Sample 1
Tuna, Sample 1
Tuna, Sample 2
Tuna, Sample 3
Tuna, Sample 4
Tuna, Sample 5
Tuna, Sample 6
Tuna, Sample 7
Tuna, Sample 8
Tuna, Sample 9
Tuna, Sample 10
Number of
Non-Detects
122
122
121
120
122
122
122
122
122
122
122
122
122
122
122
122
122
122
122
122
122
122
Number of
Detects
0
0
1
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
C-3
-------
the samples had total PCB concentrations above this tolerance limit when assuming the non-
detects are zero. The waterbodies of origin for each sample are also shown, but no clear trends
appear to connect the PCB concentrations with the waterbodies of origin.
Table C-4 shows the summary statistics from the eight catfish samples. Because several
of the samples contained much larger concentrations, suggesting a skewed distribution, both the
mean and standard deviation as well as the geometric mean and geometric standard deviation are
shown. The mean was heavily influenced by the largest concentration, suggesting the lognormal
distribution was a better approximation for the distribution. Assuming a lognormal distribution
with geometric mean and geometric standard deviation as shown, the FDA tolerance level is
above the 99th percentile of the distribution represented by the eight samples.
Table C-3. Total PCB concentrations in catfish samples
Sample
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Sample 6
Sample 7
Sample 8
Waterbody of Origin
Lake
Farm
Farm
Farm
Unknown
Atlantic
Unknown
Farm
Total PCB p.g/kg total weight
600
8.8
22
24
190
17
58
10
Lipid Fraction
0.052
0.05
0.02
0.02
0.04
0.014
0.046
0.0067
Total PCB
Hg/kg lipid
11,538
176
1,100
1,200
4,750
1,214
1,261
1,493
Table C-4. Summary of total PCB concentrations in catfish samples
represented as a lipid weight basis
Species
Catfish
Mean (|J.g/kg
lipid)
116
Standard
Deviation (|J.g/kg
lipid)
205
Geometric
Mean (|J.g/kg
lipid)
39
Geometric
Standard
Deviation
4.4
Minimum
(H-g/kg
lipid)
8.8
Maximum
(jig/kg lipid)
600
Table C-5 shows the congeners which returned detectable levels in at least one of the
catfish samples. The two samples with the highest total PCB concentrations had detectable
levels in a wide range of congeners (46 different congeners for Sample 1 and 25 different
congeners for Sample 5), while the other samples had detectable levels in relatively few
congeners (3-10).
C-4
-------
Table C-5. Congeners returning detectable levels of PCBs in catfish samples
(ug/kg lipid)
PCB Number
PCB44
PCB 49
PCB 52
PCB 56 /PCB 84 /PCB
92
PCB 60 /PCB 90 /PCB
101
PCB 66 /PCB 91
PCB 74
PCB 81
PCB 95
PCB 97
PCB 99
PCB 105
PCB 114 /PCB 146
PCB 118 /PCB 131
PCB 128
PCB 130
PCB 135
PCB 137
PCB 138
PCB 141
PCB 149
PCB 153
PCB 156
PCB 158
PCB 163
PCB 164
PCB 167
PCB 170
Sample 1
3.3
4.9
5
2.5
6.3
4.5
2.5
36
8.3
5.1
13
6.1
5.2
7.5
8.9
4.3
7.8
2.1
59
10
51
95
3.3
4.1
15
3.2
2.1
21
Sample 2
2.7
3.8
Sample 3
2.4
4.6
7.3
Sample 4
3.1
4.8
8.1
Sample 5
1.8
5
1.7
1.9
2.6
2.3
16
1.8
10
41
4.9
4.7
Sample 6
4.6
8.4
Sample 7
4.6
2
9
15
Sample 8
3.1
5
C-5
-------
Table C-5. Congeners returning detectable levels of PCBs in catfish samples
(ug/kg lipid) (continued)
PCB Number
PCB171/PCB201
PCB 172
PCB 174
PCB 177
PCB 179
PCB 180
PCB 183
PCB 187
PCB 190
PCB 194
PCB 195 /PCB 207
PCB 196
PCB 199
PCB 202
PCB 203
PCB 206
PCB 208
PCB 209
Total
Sample 1
4
4.6
13
9.7
4
67
15
40
4
11
2.5
4.9
9.9
2.4
9.1
5.3
1.5
2
600
Sample 2
2.3
8.8
Sample 3
3.8
3.4
22
Sample 4
4.4
3.7
24
Sample 5
1.6
3.5
3.2
12
4.3
14
1.4
2.3
1.7
7
4.8
29
11
190
Sample 6
3.9
17
Sample 7
6
6.9
2.9
4.8
2.7
5.1
58
Sample 8
2.1
10
Finally, Table C-6 shows the total PCB concentrations and percent lipids in all the
samples examined in the analysis. As discussed above, only the catfish, Atlantic salmon, and
blue crab species had any samples with detectable levels of PCBs. The detection limits for the
individual congeners ranged from 1.005 to 19.6 ng/kg with an average of 2.69 ng/kg. This
suggests the analytical techniques have precision which is at least a factor of 100 below the
tolerance level for each of the 124 individual congeners. These detection limits appear to be
appropriately low to resolve any total concentrations above the tolerance limit. Overall, the
particular composite samples from the species selected for PCB analysis appear to have
concentrations within the FDA tolerance level.
C-6
-------
Table C-6. Total PCBs and percent lipids in all samples
Sample
010.C
010.C
017.C
017.C
020.C
020.C
034.C
034.C
035.C
035.C
036.C
036.C
037.C
037.C
038.C
038.C
058.C
058.C
058.C
058.C
058.C
058.C
059.C
059.C
060.C
060.C
065.C
065.C
080.C
080.C
081.C
Market Name
Mackerel, Spanish
Mackerel, Spanish
Tuna
Tuna
Mackerel
Mackerel
Salmon, Atlantic
Salmon, Atlantic
Swordfish
Swordfish
Tuna
Tuna
Tuna
Tuna
Tuna
Tuna
Tuna
Tuna
Tuna
Tuna
Tuna
Tuna
Salmon, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Crab, Blue
Crab, Blue
Catfish
Catfish
Crab, Blue
Common Name
Mackerel, Spanish
Mackerel, Spanish
Tuna, Bigeye
Tuna, Bigeye
Mackerel, Atlantic
Mackerel, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Swordfish
Swordfish
Tuna, Yellowiin
Tuna, Yellowfm
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Tuna, Yellowiin
Tuna, Yellowfm
Tuna, Yellowiin
Tuna, Yellowiin
Tuna, Yellowfm
Tuna, Yellowiin
Salmon, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Crab, Blue/Softshell
Crab, Blue/Softshell
Catfish, Channel
Catfish, Channel
Crab, Blue/Hardshell
Variable
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
PERCENT LIPIDS
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
TOTAL PCB CONGENERS FISH TISSUE
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
Value (% or
US/kg)
0.82
--
0.17
--
1.3
--
8.6
--
2.1
--
0.73
--
0.64
--
0.63
--
0.54
0.54
0.54
--
--
--
10
--
11
1.5
1.7
1.5
5.2
600
1.5
Detect
Flag
Y
N
Y
N
Y
N
Y
N
Y
N
Y
N
Y
N
Y
N
Y
Y
Y
N
N
N
Y
N
Y
Y
Y
Y
Y
Y
Y
C-7
-------
Table C-6. Total PCBs and percent lipids in all samples (continued)
Sample
081.C
082.C
082.C
101.C
101.C
102.C
102.C
108.C
108.C
lll.C
lll.C
lll.C
lll.C
lll.C
lll.C
114.C
114.C
115.C
115.C
121.C
121.C
124.C
124.C
133.C
133.C
135.C
135.C
142.C
142.C
144.C
144.C
Market Name
Crab, Blue
Crab, Blue
Crab, Blue
Catfish
Catfish
Salmon, Atlantic
Salmon, Atlantic
Catfish
Catfish
Tuna
Tuna
Tuna
Tuna
Tuna
Tuna
Catfish
Catfish
Catfish
Catfish
Salmon, Atlantic
Salmon, Atlantic
Catfish
Catfish
Crab, Blue
Crab, Blue
Salmon, Atlantic
Salmon, Atlantic
Catfish
Catfish
Crab, Blue
Crab, Blue
Common Name
Crab, Blue/Hardshell
Crab, Blue/Hardshell
Crab, Blue/Hardshell
Catfish, Channel
Catfish, Channel
Salmon, Atlantic
Salmon, Atlantic
Catfish, White
Catfish, White
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Catfish, Blue
Catfish, Blue
Catfish, Blue
Catfish, Blue
Salmon, Atlantic
Salmon, Atlantic
Catfish, Channel
Catfish, Channel
Crab, Blue/Hardshell
Crab, Blue/Hardshell
Salmon, Atlantic
Salmon, Atlantic
Catfish, Blue
Catfish, Blue
Crab, Blue/Softshell
Crab, Blue/Softshell
Variable
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
PERCENT LIPIDS
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
TOTAL PCB CONGENERS FISH TISSUE
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
Value (% or
US/kg)
--
1
--
4
190
15
8.5
5
8.8
0.18
0.18
0.18
--
--
--
2
22
2
24
11
--
4.6
58
4.3
--
9.4
--
0.67
10
1.2
--
Detect
Flag
N
Y
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
Y
Y
Y
Y
Y
N
Y
Y
Y
N
Y
N
Y
Y
Y
N
-------
Table C-6. Total PCBs and percent lipids in all samples (continued)
Sample
145.C
145.C
157.C
157.C
158.C
158.C
159.C
159.C
160.C
160.C
166.C
166.C
169.C
169.C
170.C
170.C
171.C
171.C
172.C
172.C
178.C
178.C
182.C
182.C
185.C
185.C
196.C
196.C
206.C
206.C
207.C
Market Name
Crab, Blue
Crab, Blue
Tuna
Tuna
Tuna
Tuna
Tuna
Tuna
Tuna
Tuna
Catfish
Catfish
Crab, Blue
Crab, Blue
Crab, Blue
Crab, Blue
Crab, Blue
Crab, Blue
Scallop
Scallop
Salmon, Atlantic
Salmon, Atlantic
Mackerel, Spanish
Mackerel, Spanish
Salmon, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Crab, Blue
Common Name
Crab, Blue/Hardshell
Crab, Blue/Hardshell
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Tuna, Bigeye
Tuna, Yellowiin
Tuna, Yellowfm
Catfish, Channel
Catfish, Channel
Crab, Blue/Hardshell
Crab, Blue/Hardshell
Crab, Blue/Hardshell
Crab, Blue/Hardshell
Crab, Blue/Softshell
Crab, Blue/Softshell
Scallop, Sea
Scallop, Sea
Salmon, Atlantic
Salmon, Atlantic
Mackerel, Spanish
Mackerel, Spanish
Salmon, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Salmon, Atlantic
Crab, Blue/Softshell
Variable
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
Value (% or
US/kg)
3.7
--
0.25
--
0.35
--
0.54
--
0.2
--
1.4
17
4.4
--
3.6
--
1.1
--
0.11
--
9.8
--
0.27
--
10
--
6.2
--
5
--
1.3
Detect
Flag
Y
N
Y
N
Y
N
Y
N
Y
N
Y
Y
Y
N
Y
N
Y
N
Y
N
Y
N
Y
N
Y
N
Y
N
Y
N
Y
C-9
-------
Table C-6. Total PCBs and percent lipids in all samples (continued)
Sample
207.C
210.C
210.C
211.C
211.C
212.C
212.C
235.C
235.C
236.C
236.C
237.C
237.C
237.C
237.C
237.C
237.C
Market Name
Crab, Blue
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Mackerel
Common Name
Crab, Blue/Softshell
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Mackerel, Atlantic
Variable
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
PERCENT LIPIDS
PERCENT LIPIDS
PERCENT LIPIDS
TOTAL PCB CONGENERS FISH TISSUE
TOTAL PCB CONGENERS FISH TISSUE
TOTAL PCB CONGENERS FISH TISSUE
Value (% or
US/kg)
--
1.6
--
3.5
--
1.1
--
2.4
--
2.6
--
2.4
2.4
2.4
--
--
--
Detect
Flag
N
Y
N
Y
N
Y
N
Y
N
Y
N
Y
Y
Y
N
N
N
C.I. REFERENCES
U.S. FDA (2001). Fish and Fisheries Products Hazards and Controls Guidance, Third Edition.
http://www.fda.gov/Food/GuidanceComplianceRegulatoryInformation/GuidanceDocume
nts/Seafood/FishandFisheriesProductsHazardsandControlsGuide/ucm091998.htm.
C-10
-------
APPENDIX D. CALCULATIONS FOR NUMBER OF SERVINGS
Tables D-l and D-2 show the values used to determine the fish servings values shown in
Table 12 in Section 5.3. The serving values calculated using the mean and upper confidence
limits for each species are shown in columns (3) and (6), respectively. The calculated values
were truncated (i.e., rounded down) to the lower integer values shown in columns (4) and (7),
respectively. Rounding down regardless of the decimal in the calculated value is a conservative
approach to estimating the number of servings that result in intake at or below the EPA level of
concern for mercury intake.
D-l
-------
Table D-l. Estimated number offish servings per week for an adult female of child-bearing age based on means
and upper 95% confidence limits on mercury concentrations by species3
(1)
Market Name of
Species
Tuna
Swordfish
Mahi-Mahi
Spanish Mackerel
Halibut
Bluefish
Chilean Sea Bass
Pollock
Monkfish
Porgy
Croaker
Sea Bass
Lobster
Skate
Flounder
Snapper
Catfish
Cod
Whiting
Bass
Mackerel
(2)
Mean
Mercury
(mg/kg)
0.42*
0.40*
0.22*
0.15
0.15
0.15
0.13
0.13
0.11
0.098
0.084
0.075
0.069
0.060
0.051
0.049
0.044
0.031
0.028
0.025
0.022
(3)
Weekly Servings Value
Calculated using Mean
Mercury in equation
forSpW"
0.477
0.501
0.911
1.336
1.336
1.336
1.542
1.542
1.822
2.045
2.386
2.673
2.905
3.341
3.930
4.091
4.555
6.466
7.159
8.018
9.111
(4)
Number of servings per
Week using Mean
Mercury that result in
intake at or below the
EPA level of concern
0
0
0
1
1
1
1
1
1
2
2
2
2
3
3
4
4
6
7
8
9
(5)
95% Upper
Confidence Limit
on Mean Mercury
(mg/kg)
0.55
0.59
NA
0.20
NA
0.17
NA
0.15
0.13
0.12
0.10
0.088
NA
0.078
0.065
0.060
0.061
0.038
0.043
0.047
0.028
(6)
Weekly Servings Value
Calculated using 95%
Upper Confidence Limit
on Mean Mercury in
equation for SplV*1
0.364
0.340
NA
1.002
NA
1.179
NA
1.336
1.542
1.670
2.004
2.278
NA
2.570
3.084
3.341
3.286
5.275
4.661
4.265
7.159
(7)
Number of servings per Week
using 95% Upper Confidence
Limit on Mean Mercury that
result in intake at or below the
EPA level of concern
0
0
NA
1
1
1
NA
1
1
1
2
2
NA
2
3
3
3
5
4
4
7
o
to
-------
Table D-l. Estimated number of fish servings per week for an adult female of child-bearing age based on means
and upper 95% confidence limits on mercury concentrations by species" (continued)
(1)
Market Name of
Species
Ocean Perch
Herring
Oyster
Blue Crab
Tilapia
Squid
Mussel
Rainbow Trout
Clam
Atlantic Salmon
Scallop
Shrimp
(2)
Mean
Mercury
(mg/kg)
0.022
0.022
0.015
0.015
0.014
0.014
0.012
0.012
0.0081
0.0081
0.0055
0.0054
(3)
Weekly Servings Value
Calculated using Mean
Mercury in equation
forSpW*
9.111
9.111
13.363
13.363
14.317
14.317
16.703
16.703
24.746
24.746
36.444
37.119
(4)
Number of servings per
Week using Mean
Mercury that result in
intake at or below the
EPA level of concern
9
9
13
13
14
14
16
16
24
24
36
37
(5)
95% Upper
Confidence Limit
on Mean Mercury
(mg/kg)
NA
NA
0.025
0.021
0.022
0.017
0.017
NA
0.012
0.012
0.0071
0.0073
(6)
Weekly Servings Value
Calculated using 95%
Upper Confidence Limit
on Mean Mercury in
equation for SpW^
NA
NA
8.018
9.545
9.111
11.791
11.791
NA
16.703
16.703
28.231
27.458
(7)
Number of servings per Week
using 95% Upper Confidence
Limit on Mean Mercury that
result in intake at or below the
EPA level of concern
NA
NA
8
9
9
11
11
NA
16
16
28
27
aServing values calculated using the following exposure assumptions: Serving size = 8 oz offish fresh weight, Adult female weight = 65 kg, RID for methyl mercury = 1 x 10"4
mg/kg-day (U.S. EPA IRIS database), and the person consumes only the one type offish or shellfish.
^Weekly value calculated using the equation for SpW in Section 5.3.
*These concentrations yield values indicating less than one serving a week results in intake at or below the EPA level of concern. Servings per 30 day month of these species that
result in intake at or below the EPA level of concern are provided in Table D-2.
-------
Table D-2. Estimated number of fish servings per 30 day month for an adult female of child-bearing age based
on means and upper 95% confidence limits on mercury concentrations for three high mercury species
(1)
Market
Name of
Species
Tuna
Swordfish
Mahi-Mahi
(2)
Mean
Mercury
(mg/kg)
0.42
0.40
0.22
(3)
Servings Value per
30 Day MONTH
Calculated using
Mean Mercury °
2.045
2.148
3.905
(4)
Number of servings per 30
Day MONTH using Mean
Mercury that result in intake
at or below the EPA level of
concern
2
2
3
(5)
95% Upper
Confidence Limit
on Mean Mercury
(mg/kg)
0.55
0.59
NA
(6)
Servings Value per 30 Day
MONTH Calculated using
95% Upper Confidence
Limit on Mean Mercury3
1.562
1.456
NA
(7)
Number of servings per 30 Day
MONTH using 95% Upper
Confidence Limit on Mean Mercury
that result in intake at or below the
EPA level of concern
1
1
NA
aValue per 30 day month calculated using the equation for SpW in Section 5.3, with 30 days rather than 7.
-------
APPENDIX E. EXTERNAL PEER REVIEW OF EPA'S DRAFT REPORT, FISH
TISSUE ANALYSIS FOR MERCURY AND PCBS FROM A NEW YORK CITY
COMMERCIAL FISH/SEAFOOD MARKET AND EPA'S RESPONSE TO COMMENTS
E-l
-------
EXTERNAL PEER REVIEW OF EPA'S DRAFT REPORT,
FISH TISSUE ANALYSIS FOR MERCURY AND PCBs FROM A
NEW YORK CITY COMMERCIAL FISH/SEAFOOD MARKET
Contract No.: EP-C-07-024
Task Order 114
Submitted to:
Cheryl Itkin
U.S. Environmental Protection Agency
Office of Research and Development
National Center for Environmental Assessment
1200 Pennsylvania Avenue, NW
Washington, DC 20460
Submitted by:
Eastern Research Group, Inc.
110 Hartwell Avenue
Lexington, MA 02421-3136
July 25, 2011
Printed on Recycled Paper
-------
-------
QUALITY NARRATIVE STATEMENT
ERG selected reviewers according to selection criteria provided by EPA. EPA confirmed that the
scientific credentials of the reviewers proposed by ERG fulfilled EPA's selection criteria. Reviewers
conducted the review according to a charge prepared by EPA and instructions prepared by ERG. ERG
checked the reviewers' written comments to ensure that each reviewer had provided a substantial response
to each charge question (or that the reviewer had indicated that any question[s] not responded to was
outside the reviewer's area of expertise). Since this is an independent external review, ERG did not edit
the reviewers' comments in any way, but rather transmitted them unaltered to EPA. ERG did, however,
format reviewers' comments as needed for consistency in this final peer review summary report.
-------
Contents
Responses to Charge Questions 1
1. Please comment on the organization and clarity of the report 3
2. Is the data adequate for meeting the objectives of this study? 11
3. Is the select! on of the fish species adequate for this study? 13
4. Is the use and presentation of the descriptive statistics appropriate? 16
5. Please comment on the data summary in Table 14. "Estimated Fish Servings
per Week for an Adult Female of Child-bearing age based on Means and
Upper 95% Confidence Limits on Mean Mercury Concentrations by
Species". This table will draw a lot of attention. Is the table clear and does it
provide the appropriate message? 20
6. Were the analytic methods for obtaining mercury and PCB concentrations in
fish tissue appropriate for this study? 22
7. Given that the fish specimens were obtained from commercial venders and
some specimens were not whole fish, the EPA implemented DNA analysis
for verification of species. Please comment on this approach and how the
information is presented in the appendix of the report 24
Appendix A: Individual Reviewer Comments A-l
Henry A. Anderson, M.D A-3
Gary A. Pascoe, Ph.D., DABT A-ll
AlanH. Stern, Dr.P.H., D.A.B.T A-19
-------
Responses to Charge Questions
-------
-------
Responses to Charge Questions
1. Please comment on the organization and clarity of the report.
Reviewer
Comments
Response to Comments
Anderson
The organization of the report is appropriate and easy to
follow. There are a number of approaches that would help
with clarity. The report is quite long and presents many
tables and figures which are very useful. This makes the
executive summary become particularly important. Right
now the executive summary is pretty cursory and the key
findings of the report need to be highlighted in some kind
of order. Having them as bullets would improve the
clarity. It might also help if the key findings were
highlighted in each section of the report at the front.
These could be bullets but listing them and then having
the text explain and elaborate would help. For instance,
there is a lot of interest in farmed vs wild fish. This is
mentioned and information provided in the text and there
is not much to say since only three species were farm
raised, and only one of those a fin fish. But it might be
mentioned that no mercury was detected in the wild
Atlantic salmon but found in half of the farmed salmon,
even though the levels were very low. Farm vs wild is
only included as part of "water body" discussion. If the
authors go through each section and select what they view
as the key findings, list them in the sections and then carry
them into the executive summary that would help the
reader focus on key findings.
The impetus for the study was the New York City blood
mercury report. This study and its findings should
probably be summarized a bit better and the fish
consumption rates mentioned as well as the mercury
prevalence's in the general population and other
associated factors. The question that this report does not
address is whether these fish monitoring data presented
can explain the blood mercury distributions seen? It
would help if somehow the authors could translate the
various "action levels" presented into what are their blood
mercury level equivalents. Or at least discuss how the
meal rates reported in Table 14a might translate into
blood mercury.
The Executive Summary (ES) has been
re-written to comprehensively address
all reviewer comments related to the ES.
A new section (3.4) has been added that
qualitatively discusses the wild vs
farmed results.
The suggestion to translate "action
levels" and species specific seafood
meal recommendations in the report
(Tables 14 A and B [now Tables 12 and
13]) into an equivalent blood mercury
(Hg) concentration is an interesting and
insightful thought; however, it would
require a significant biokinetic analysis
which is beyond the stated goals of this
study. It is a subject worth consideration
for a future spin-off paper.
What does the NY 5 ug/L blood translate into as fish
The NYCHANES study notes that NYC
residents have both higher blood Hg
-------
Responses to Charge Questions
contaminant level and meals per week? Will the Table
14a result in blood mercury less than the NY 5 ug/L
value? It is somewhat counter intuitive that none of the
composites exceed the FDA action level, and very few
the EPA Rfd based level, yet 72% of the NYC Chinese
exceed the NY reportable level of 5 ug/L mercury in
blood. If these "action levels" are to protect everyone at
unlimited fish consumption levels, then these "action
levels" don't seem to be protective as far as the NY
reportable level is concerned. Some discussion of the
basis for the NY reportable level is needed and how it
relates to the report's findings. In the executive summary
it is mentioned that the NYC fish samples appear to be
somewhat lower than the FDA comparisons. That triggers
the question of then why do NYC residents have blood
mercury levels three times those of the rest of the
country? I would suggest that the testing samples are
qualitatively similar, but doing a statistical comparison is
problematic because of the different sampling structure.
In the introduction it is mentioned that the "action levels"
used will be discussed further in the report. But all that is
included is a table and references. There is no discussion
of how they were derived, why they are different and how
they are intended to be used. This is a bit of apples vs
oranges. What is the level offish consumption that each
assumes? The challenge in understanding this report is
how to arrive at a "dose" by navigating between fish
tissue concentration, fish consumption rates over a
selected time period and the blood mercury levels seen in
the NYC study. It appears simple but is complex.
levels than national levels and higher
fish consumption rates as well. Text has
been added to the introduction section of
Section 1 that articulates this
information.
Derivation of "action levels" is beyond
the scope of this study.
Pascoe
The report is well organized and clearly written, with a
few exceptions noted below; nonetheless, it should be
easily understandable by health professionals and the
interested public. The following are editorial comments
that either need addressing or may help with clarity.
These comments focus on the Executive Summary, which
should be both comprehensive yet simplified enough with
explanations of technical issues in order to be sufficiently
understandable to someone unfamiliar with those issues.
Page 1, Executive Summary, first line - Because some
states use the same or similar regulatory agency name,
EPA should be referred to as the U.S. Environmental
Protection Agency upon first use in the Executive
Summary and the main body of the report.
No response needed.
The Executive Summary (ES) has been
re-written to comprehensively address
all reviewer comments related to the ES.
-------
Responses to Charge Questions
Page 1, Executive Summary, first paragraph - The goal of
the study to support the NYC public health message
should be clarified with more detail. The development of
a large seafood mercury database to provide support for
the City's message on public health is highly admirable,
and the readers could be more informed of how the data
were actually used in formulating the health messages,
and what those messages are. The final tables in the
report provide recommendations on meals per month that
sensitive members of the population (e.g., women of
child bearing age) can consume of different types offish
from the Commercial Market, and the development of
these recommendations using the mercury seafood data
that the study collected should be more strongly linked to
the NYC public health messages.
Page 1, Executive Summary, second paragraph - The
derivation and meaning of the term "reportable level" is
not explained, and may have different meanings to a
chemist and a health professional or someone not versed
in those fields. The term should be clarified as to whether
it is a health-based criterion or a departmental advisory,
or an enforceable standard, or whether it refers to
concentrations that should be reported to a regulatory
agency, and if so the basis for the requirement and the
concentration of mercury should be noted.
Page 1, Executive Summary, third paragraph, fourth
sentence - A comma appears to be missing.
Page 2, Executive Summary, third paragraph - The first
sentence appears to state that the report estimates the
amount of seafood adult women of child bearing age
consume, from which the intake of mercury is estimated.
This is an incorrect description of the process that is
actually used in the report, in that the amount of seafood
consumed by adult women of child bearing age is not
itself estimated, but instead what is estimated is the
amount of seafood consumption that corresponds to a
level of concern for exposure to mercury. If the report
intended to estimate the amount of seafood that a
population or subgroup of a population consumes, an
entirely different type of consumption study would be
needed. The amount of seafood consumption considered
to correspond to a level of concern is independent of
actual seafood consumption rates. This distinction is
critical primarily because some subpopulations will
consume more seafood than others, which is noted at the
The Executive Summary (ES) has been
re-written to comprehensively address
all reviewer comments related to the ES.
Beyond widely deciminating this report,
the onus will be on the NYCDOHMH to
incorpoarate the findings of the report
into their public health message of
seafood consumtion.
The Executive Summary (ES) has been
re-written to comprehensively address
all reviewer comments related to the ES.
The authors disagree with the reviewers
comment re: the message communicated
in the first sentence of paragraph 3, ES.
There is no intent in the study to
estimate the amount of seafood women
of child bearing age consume. As
subsequently noted by the reviewer, the
goal of the study is to estimate the
allowable number of seafood meals (on
a species-specific basis) for women of
child bearing age.
-------
Responses to Charge Questions
end of the report, and descriptions of seafood
consumptions by subpopulations need to be sensitive to
such cultural differences. The fourth sentence of this
paragraph also alludes to a determination of the amount
of seafood that a subpopulation consumes - the
permissible daily intake of methyl mercury was not
actually compared with the amount of methyl mercury
intake from fish ingestion, but rather was used to identify
the permissible amount of ingestion of fish that contain
varying levels of methyl mercury.
Page 2, Executive Summary, fourth paragraph and
Section 6 - The report should attempt to provide some
explanation of why the NYC CM data were so much
lower than the FDA data on mercury in fish tissue, if
there is a known or suspected reason.
Page 3, Executive Summary, first full paragraph -
Additional summary of the rationale for the use of the
barcoding would be appropriate here. A rationale is
alluded to at the beginning of Appendix B which
discusses the problems of taxonomic identification of
dead fish. It should be emphasized in the Executive
Summary that the barcoding was considered useful
because the sources of the fish specimens consisted of
different commercial fishers with uncertain quality of
handling of the specimens, and the sources were not from
a scientific survey and collection from the water bodies
from which NYC fish are typically caught. Additionally,
there was concern that the handling of the fish specimens
by the commercial fishers could result in enough damage
to the fish exterior to make it difficult to identify at a
species level, as well as the reasons stated in Appendix B.
In addition, more summary of the utility of the barcoding
to the study should be added to the end of the Executive
Summary. Specifically, the first full paragraph of Page
113 mentions that disparities between taxonomic
identification and barcoding also resulted in re-
examination of the photos of collected specimens by EPA
to determine whether they had been mis-identified.
Page 31, Section 5.2, third paragraph - This text
discusses the comparison of total mercury concentrations
in fish with the EPA action level for methyl mercury. The
reason given as to why EPA considers it acceptable to
compare total mercury concentrations with the action
With the exception of sword fish (which
had a small sample size) and bluefish,
the study results were not neccesarily
"so much lower" than the FDA data. All
species with the exception of swordfish
and bluefish had sample means that were
within one standard deviation of FDA
sample means. The results are
appropriately described in the Executive
Summary (i.e., "tended to be lower than
FDA").
Revisions were made as per the
reviewer's comment, (all but the first
sentence of Section 5.2, paragraph 3 was
removed).
-------
Responses to Charge Questions
level for methyl mercury is not entirely correct. EPA
(2009) does not state that the comparison is acceptable
because the molecular weight of mercury is close to that
of methyl mercury, but because most if not sometimes all
of the total mercury in fish tissue is actually methyl
mercury, as stated in the text of the subsequent paragraph
(fourth paragraph of Section 5.2). The text of the third
paragraph needs to be changed to reflect the actual
reasoning contained in EPe (2009) for the comparison.
Page 34 - The reference to Figure 10 in Appendix A
should be Figure 3.
Page 37 - The reference to Table D-la appears to refer to
Table 14.
Page 43, Section 6 has a formatting problem.
Page 54 - Two references for EPA 2009 are provided;
however, the second reference of EPA (2009) Guidance
for Implementing the January 2001 Methylmercury Water
Quality Criterion has been superseded by an updated
document: USEPA. 2010. Guidance for Implementing the
January 2001 Methylmercury Water Quality Criterion.
April. EPA 823-R-10-001.
http://water.epa.gov/scitech/swguidance/standards/criteria
/
aqlife/pollutants/methylmercurv/upload/mercurv2010.pdf
Editor/corrected.
Editor/corrected.
Editor/corrected.
Stern
In general, the report is well organized and clearly
written. There are some specific points where the text is
unclear. These are noted in my specific comments below.
The one significant exception to this, as discussed in
detail below, is that the full objectives of the report as
evinced by the analyses in the report, itself, are not
clearly stated. Notwithstanding the overall logical
organization and general clarity of the text, however, the
report seems to reflect an inconsistent level of technical
detail. Some sections are technically quite detailed and
appear not to be intended to be generally accessible to the
lay readier such as, for example, Section 4, "Analysis of
Measurement Variability" and Appendix B, the
explanation DNA "barcode" analysis. While Section 5,
"Comparison of CM Data to Risk Metrics" is much more
accessible. Some thought could be given to the intended
audience for this report.
The report is intended to reach a wide
audience of readers: from the lay public
for end-user advice on selection of
species and the frequency of seafood
meals, to health professionals for
crafting public health messages and for
general research purposes. Parts of the
report's content are necessarily more
complex (e.g., DNA barcoding,
statistical analysis). To the extent
feasible, the report has employed
simplified text in communicating is
primary goal: recommendations on types
and frequencies of seafood meals.
The report compares the NYC commercial market Hg
The authors agree that an analysis of
-------
Responses to Charge Questions
concentrations by species to the FDA database of Hg
concentration by species and gives the implications for
consumption frequency advisories relative to the EPA
RfD. Further, the report finds that, in general, the
concentrations from the NYC commercial market
samples are significantly lower on a species-by-species
basis. The obvious next logical step is to compare the
current FDA/EPA advisory specifics to the meal
frequencies derived in the report (Table 14). This is not
done and is conspicuous by its absence.
Specific Comments
Pg. 11, par. 2 - The text is unclear as to how duplicates
and replicates were treated in terms of the value for each
composite sample that was actually entered into the
overall statistical analysis.
Pg 12, par. 2 - The treatment of non-detects appears to
have little overall impact on the estimation of the average
Hg concentration across species and on the use of these
data for fish consumption advisories (because the non-
detects are in the species with the lowest Hg
concentrations). It should, however, be noted for the sake
of completeness that the comparison between C = ND/2
and C = ND is a biased comparison since it assumes that
the true value of the concentration in the case of a non-
detect can only be larger than the theoretical average (i.e.,
ND/2). A more balanced comparison would be among C =
ND/4, C = ND/2 and C = ND.
current FDA advisories to seafood meal
frequencies cited in this report would be
informative, but such an exercise is
beyond the stated goals of this study.
The paragraph that starts at the bottom
of page 10 and continues to the top of
page 11 describes how duplicate and
replicate analyses on the same composite
were handled. The last sentence on page
10 reads as follows: "For the purposes of
statistical analysis, the duplicates or the
replicates were averaged to obtain a
single Hg concentration for the
respective composite sample."
This text is now on the bottom of page 3
of the Final draft 1. The description here
and on the top of page 4 should clarify
the issue raised in the comment.
(see also the comment by Pascoe re: use
of ProUCL in response to question #4).
The use of C = ND/2 and C = ND was
intended to explore the sensitivity of the
descriptive statistics results to legitimate
alternative values for the non-detects and
the results show, as the reviewer notes,
little overall impact on the averages.
There was however, no intent to bias the
results or to perform an exhaustive
analysis of the effects of possible values
for measured values reported as non-
detect. In fact, the theoretical average of
the non-detects is unknown and the
distribution of the non-detects is
unknown (it could, for example be
highly skewed) and non-detect values
are not reported. In any case, C = ND/2
is more likely an approximation to the
median non-detect value although this
also is not known.
The text referred to in the comment is
not now on page 12 of the Final draftl.
There is a considerable amount of
discussion with tables in Sections 2, 3,
-------
Responses to Charge Questions
Pg. 21, par. 2 - The reference to large R2 values resulting
from "large separations in data points" sound like the
effect of influential/outlying observations. If the concern
is with such observations unduly influencing the
correlation, were non-parametric correlations considered?
Pg. 22 (equation) - The equation is not comprehensible as
printed although it can be decoded from the variable
definitions below.
Pg. 31, par. 3 - ... and includes attributes of the level
important to its meaning or implementation. " This
language is unclear.
Pg. 37, par. 3 - "Table D-la" The primary reference here
should probably be to Table 14.
and 4 that deals with the non-detects
There was a general tendency for fish
size to be positively related to mercury
concentrations regardless of the R2
values. The distortion in the R2 values
were generally due to large separations
in clusters of data points (i.e., not
necessarily the effect of individual
influential observations) resulting in a
graphical pattern that cast doubt on the
validity of the usual linear regression
type of relationship. Non-parametric
correlations were not considered.
The discussion [of the R2 values] is
adequate. Generally the problem with
the R2 values is not due to individual
influential observations, which is clear
from the text and the response to the
comment. On-parametric correlations
would not remedy this problem.
Editor reformat/corrected.
Subscript on x; corrected which should
help to clarify.
Text was revised to read: "... and
includes ancillary descriptive
information."
Agree (Editor revised). Editor: this
paragraph should begin with the words
"This table.. ."/corrected.
Editor to revise/corrected.
-------
Responses to Charge Questions
"Converting to meals per 30-day month as shown in
Table 14" The reference here should be to Table 14b.
"Using the conservative estimate ofHg concentration...
also should not be eaten weekly. " The issue at this point
in the text is monthly consumption. Weekly consumption
was already addressed.
Pg. 43, par. 2 - The format needs to be edited.
Par. 3 - The basis for CM-FDA matching should be
clarified.
Table 15-1 don't understand why the species detail in the
CM data present in Table 2 was not carried forward here.
Pg. 45, par. 3 - "...the mean CM concentration...the only
exception is bluefish" Swordfish also appears to be an
exception.
Text makes sense as worded - no
revision needed.
Note to editor: remove the comma from
the next to last line of paragraph 1, Pg
43/corrected.
Editor revised/corrected.
The basis for CM-FDA matching is
sufficiently descriptive - no revision
needed.
Note to editor: Pg 43 paragraph 3 should
begin with the words: "Table 15..."
Corrected to Table 14.
The purpose of Table 15 (now Table 14)
is to compare CM market names with
FDA monitoring names. Adding the
analytical data from Table 2 would make
the table unwieldy and cluttered - no
revision needed.
Agree - Swordfish is also an exception.
Note to editor: The reference to Figure 3
in this paragraph is misplaced. The
paragraph needs to be revised to
reference Table 16/corrected.
Text refers to two additional species
(other than predetermined group of 5)
that were analyzed for PCBs as well as
Hg. No revision needed.
No revision to text indicated.
10
-------
Responses to Charge Questions
Pg. 117, par. 1 - "Two samples...selected only for
mercury analysis were also analyzed for PCBs" I don't
understand this.
Pg. 119, par. 2 - "The waterbodies of origin for each
sample are also shown, but no clear trends appear to
connect the PCB concentrations with the waterbodies of
origin." Another way to look at the data in Table C-15 is
that the farm-raised catfish were uniformly low in PCBs.
2. Is the data adequate for meeting the objectives of this study?
Reviewer
Anderson
Pascoe
Comments
It is unclear whether the data meets the object of measuring
mercury in commonly consumed seafood species in New
York. Certainly the data represents the fresh seafood
consumed in the city. What is not addressed is what
percentage of seafood consumed is consumed fresh. If there is
any data on total seafood from the city or other surveys,
reporting that would help place this data in perspective. The
data is certainly important and very useful. The data and
methods descriptions are very good and comprehensive. It is
easy to understand what was done and mostly why. It should
be helpful to the NYC "Eat Fish, Choose Wisely" project.
The PCB information, although an ancillary activity is very
useful and provides a great deal of information that augments
the paucity of commercial fish PCB data. Most of the PCB
mention is relegated to the appendix C except for brief
mention in the executive summary. The critical statement in
the executive summary is that the PCBs appear within the
FDA tolerance level. For mercury the authors selected several
"action levels" to compare to not just the FDA tolerance.
Many states, as well as the EPA have "action levels" for
PCBs which are, as with mercury, considerably lower than
what FDA uses. While the PCBs are all quite low, including
the EPA RfD or some of the states' or the Great Lakes
Protocol value would be informative. Several states have
catfish on their advisories. The highest catfish with PCB
appears to be a wild caught lake fish. So for comparative
purposes, having the low to high "action levels" used by
various entities for PCB would be appropriate.
The data collected by NYC and EPA are fully adequate for
meeting the primary objective of the study. Based on the
objective of analyzing mercury in the different seafood types
Response to Comments
As the report notes, and the
reviewer acknowledges, the limited
PCB analysis was an ancillary
activity to the main goal of the
report. Accordingly, in keeping with
the level of effort for this ancillary
activity and the overall scope of the
report, the text will provide only the
FDA Tolerance Level for PCBs in
fish.
No response needed.
11
-------
Responses to Charge Questions
that are consumed by people in NYC, the report sufficiently
characterized mercury in a large suite of seafood types,
covering multiple trophic levels and multiple phylogenetic
levels. The study collected sufficient numbers of samples to
enable statistical analyses among the different categories of
seafood for total mercury concentrations, and among species
within some of the categories.
In Appendix C, the data collected on PCB concentrations in
fish specimens were insufficient to perform a similar level of
statistical analysis as was performed for mercury, but the data
collection was not a primary objective of the study and the
data still provide useful information on levels of PCBs that
the general fish consuming public might be exposed to. The
discussion on page 122 about the adequacy of the detection
limits could be expanded to include comparison of total PCBs
based on assuming non-detected congener concentrations at
the detection limit rather than as zero, which would appear to
still result in the concentration of total PCBs based on
congeners to be below the FDA action level.
The report mentions the availability of blood levels of
mercury for NYC residents. If studies have been performed
that link those blood levels to patterns of seafood
consumption and types offish, particularly for ethnic or
cultural subpopulations, some discussion should be
considered in the report that would more fully link those
studies with the objectives of the CM study.
As discussed in the report, due to
the many PCB congeners analyzed,
assigning a value of l/i the detection
limit for non detects "could severely
overestimate the total PCB
concentration." Assigning a value
of zero to non detects "will tend to
underestimate the total PCB
concentration but likely represents
the method with the smallest bias."
Accordingly, no revision indicated.
The NYC HANES study
(McKelvey et al., 2007) reported on
the patterns of seafood consumption
in ethnic/racial subpopulations
within NYC noting that New
Yorkers of Asian descent had both
the highest blood Hg levels and the
highest self reported fish
consumption. Text was added to
Section 1 describing blood Hg
levels and seafood consumption
patterns.
Stern
The specific objectives of the study do not appear to be stated
in the report. This is a shortcoming that should be rectified.
The first paragraph of Section 1 states that study was
"...undertaken to measure mercury (Hg) concentration in
composite samples from seafood species most commonly
consumed by New York City residents as represented by
specimens obtained from a commercial market." Based on
the sections of the report that compare the NY City market
Hg concentrations to those from the same species in the FDA
database and based on the calculation of the meal frequency
The objective of the study is to give
New York City residents
information (Hg concentration) on
the species of seafood most
commonly consumed by area
residents so they can make informed
choices on the type and frequency
of seafood meals that will minimize
(i.e., maintain exposure below the
RfD) their Hg exposure. As noted
previously, a larger exercise to
12
-------
Responses to Charge Questions
by species that would exceed the RfD, it is clear that the
objectives of the report go beyond simply measuring Hg
concentration in common commercial fish. The apparent
objective of the study is to assess the currency and accuracy
of the existing FDA/EPA mercury advisory for commercial
fish and to recommend possible adjustments based on the
more up-to-date NY City commercial market data. To the
extent that the objectives are focused on the FDA database
and on the FDA/EPA advice relative to the NYC wholesale
market per se, the study appears to meet those implied
objectives both in terms of sampling and statistical analysis.
However, given the absence of clear statement of objectives,
it is not clear to what extent EPA intends the data from the
NYC commercial market to be representative of commercial
fish Hg data nationwide. Section 1 does state the goal of the
study was to measure Hg levels in fish commonly consumed
by NYC residents. Nonetheless, the study clearly has broader
implications. The NYC commercial fish market is a major
intake and distribution point for commercial fish in the
eastern U.S. and the NYC wholesale market receives fish
caught in global waters. Nonetheless, there are other
wholesale commercial fish markets in the eastern U.S. and
there are other wholesale markets that serve other parts of the
U.S. Thus, it is unclear to what extent EPA intends to
generalize the findings of this study to the U.S. as a whole. A
clear statement of the objectives of this study should include a
discussion of the intent of its focus on the NYC market and
the extent to which EPA believes that data from the NYC
market can and should be used to generalize to commercial
fish nationwide.
evaluate the overall adequacy of
FDA regulations/advisories re: Hg
in fish is not the intended goal of the
study and beyond its scope. That
said, The Executive Summary and
introduction to Chapter 1 has been
revised to more clearly state the
specific aims of the study.
As noted above, study objectives
have been clarified. The study grew
out of the results from the NYC
HANES (McKelvey et al., 2007)
study which reported that relative to
national data New York City adults
had higher blood Hg concentration
and a greater frequency of seafood
meals. The results of the study are
not intended to be generalizable to
the nation as a whole.
3. Is the selection of the fish species adequate for this study?
Reviewer
Comments
Response to Comments
Anderson
The selection offish species is adequate for the general NYC
population who purchase fresh fish. The sample sizes while
not large are sufficient. While analyses of single fish would
have been preferred simply because that is what is most
descriptive of consumer activity, the use of composites is
acceptable. This does make it difficult to directly compare
results from other studies. But, the statistical analyses
provided show that the findings are consistent with what has
been found in other studies. It would be helpful to have a bit
more explanation of how the use of composites rather than the
same number of samples but of individual fish strengthens the
The basic unit of analysis in the
study is a composite sample offish
tissue by species. From the first
paragraph in the report (page 4):
"The New York City Commercial
Market (CM) Seafood Study was
undertaken to measure mercury
(Hg) concentration in composite
samples from seafood species most
commonly consumed by New York
City residents as represented by
specimens obtained from a
13
-------
Responses to Charge Questions
findings. No major commercial species appears to be missing.
commercial market." "Each
composite sample was formed by
mixing tissue from a number
individual fish specimens into a
combined amalgamated sample. The
formation of the composite sample
is, in effect, a physical averaging of
the individual tissue samples and
the result of a single measurement
on the composite sample is an
estimate of the average of the
specimens in the sample. Composite
sample analysis is a well established
mechanism for cost effective
estimation of means of
environmental samples that has a
history of use in fish tissue analysis
(e.g., Procedures for Formation of
Composite Samples from
Segmented Populations, Fabrizio et
al., Environ. Sci. Techno!., 1995, 29
(5), 1137-1144, Gilbert, R.O.,
Statistical Methods for
Environmental Pollution
Monitoring, 1987.)." The use of
estimates of mean concentrations
(derived from measurements on
composite samples) supports the
objective of assessing consumer
behavior over time which is
indicative of chronic exposure.
Sampling of frozen, canned and
dried fish sold in Asian markets and
biomonitoring of blood mercury
data were not in the scope of this
study.
This response and the document are
responsive to the comment about
composite samples. Questions
about the Chinese or Asian intake of
different species are outside the
scope of our study. It has been
included in the response. It could be
emphasized more as some readers
may have a similar reaction.
14
-------
Responses to Charge Questions
Given that the report introduction and background emphasizes
the blood mercury levels in the NYC Chinese, what is missing
is any sampling of the imported frozen, canned and dried fish
that are sold in Chinese/oriental groceries. Some mention of
any published studies showing that these fish are similar in
contaminant concentration than other fish would help. It is
unclear how much of the fish in Chinese diets come from the
species sampled in this study. The lack of ethnic specific
commercial fish should at least be discussed and might help
explain the biomonitoring observations rather than just that
for this population frequency of consumption probably
accounts for the differences. At least some discussion or at
least the recognition of the possible impact of ethnically
preferred fish needs to be mentioned. As mentioned in
response to #1, it would be good to include the biomonitoring
blood mercury data for the general NYC population and not
just the Chinese. This study is not ethnic focused.
The original study design, in
partnership with the NYCDOHMH,
was to do city -wide sampling of
seafood with oversampling in
predominantly Asian communities
to capture cultural differences in
species consumption. Ultimately,
the studies were bifurcated with the
NYCDOHMH focusing exclusively
on markets in Asian communities.
(see McKelvey, Chang, et al., 2010)
Pascoe
The species selected for the study address multiple types of
seafood that are typically consumed by the targeted audience,
with the variety ranging from shellfish to squid to various
sizes and types of finfish. Since the objective of the study
focused on the types of seafood that consumers would
typically eat from a commercial market, the variety in the
types of seafood that were collected from the market was
appropriate and adequate. Because the study was able to
analyze fish and shellfish from multiple trophic and
phylogenetic levels, it exceeded the needs of the study with
regards to variety in seafood types.
No response needed.
Stern
As above, it is not possible to tell the extent to which the
selection offish species is adequate without knowing the
overall objective of the report and more specifically, without
knowing the extent to which EPA intends to generalize from
the data collected from the NYC commercial market. The
stated intent of the collection process was to obtain a
representative sample of the fish most commonly consumed
by NYC residents. The report briefly cites several databases
(NMFS fish landings data; "Fish availability in
supermarkets... in New Jersey" (Burger et al., 2004, but not
specifically identified as such in the report); and the "Seafood
Products Matrix" (not further defined)) as the basis for the
selection of fish obtained from the NYC market. Given the
level of detail in other parts of the report, it is surprising that
no information was provided as to how these databases were
actually used to determine the market sampling strategy. This
information should be included in the report. The list of fish
included in the study appears to be appropriate and does, in
15
-------
Responses to Charge Questions
my subjective assessment, include all or most fish commonly
found in NYC markets. Nonetheless, given the nature of the
study, documentation of the process seems essential.
4. Is the use and presentation of the descriptive statistics appropriate?
Reviewer
Comments
Response to Comments
Anderson
The use of descriptive statistics is appropriate and when
reported are presented with appropriate caveats about sample
size etc and explanations of why certain procedures were or
were not used. This was very helpful.
The description of the analyses of variability is quite good and
comprehensive. It is probably too detailed for most readers
and all the formulae are quite intimidating. These details
might be better included as an appendix and just the results
and general methods discussed in the main text. What is also
needed is a discussion of the a priori acceptable level of
variability or what variability is scientifically acceptable. A
clear set of concluding statements are needed. These need to
indicate that the laboratory analyses and sample processing
methods were assessed for consistency and reliability and
found to meet accepted scientific standards. Actually the
minimal variability found is quite impressive and shows the
rigorous quality control that was utilized.
I don't think it is appropriate to add the replicates and
duplicates to the primary sample and average them and then
assign that value to the sample. This really messes up the
variability since they are no longer individual composite
results but means. I doubt it makes much practical difference,
but I would prefer that the primary sample result be the only
result used in the overall analyses. Then all the composite
sample results are truly comparable. The replicates and
duplicates should be used only for what they were intended to
be used for - assessing variability. The report quite effectively
presents the impact of various analytic decisions such as how
to handle the non-detects - showing that the differences in
methods make little impact on conclusions. The same "with
and without" approach could be done for the
duplicate/replicate averaging impact. Rather than redo all the
tables without the replicate/duplicate data included, you could
see if it makes a difference and if it does not, simply state that
to be the case. Then explain why you believe having some
values as means and others not is more robust and statistically
appropriate and that is why you chose to "use all the data".
But whether it is appropriate and allows use of the statistical
procedures used in other tables, I will leave to a statistician to
No response needed.
An appendix would be appropriate
but not necessary especially if the
comment document is included.
Added concluding statements - text
to the document on page 19.
Thanks again to Anderson for
pointing out that the minimal
variability found and the demo of
good quality control.
Composite samples are a physical
average of all the individual
specimens in the composite. A
single analysis of the composite
sample is by definition an estimate
of the mean of all the individual
specimens in the composite and
thus composite sample analysis is a
cost effective mechanism for
obtaining an estimate of the mean
of a number of individual
specimens. Replicate and
duplicate analyses provide
valuable information on the
content of the composite and
assigning the value of the average
of replicate and duplicate analyses
to the composite is a standard
practice. (See, for example:
http ://water. epa. gov/scitech/swguid
ance/fishshellfish/techguidance/ris
k/upload/2009_04_23_fish_advice
volume 1 vlcover.pdf
http://www.epa.gov/reg3hwmd7ris
k/eco/faqs/composite.htm.')
16
-------
Responses to Charge Questions
decide.
Individual measurements on
composite samples are, by
definition, estimates of the mean of
the individual specimens in the
composite and thus averages of
composite sample replicates and
duplicates will have somewhat
smaller variance compared to
individual measurements but the
assertion that "This really messes
up the variability since they are no
longer individual composite results
but means" is not correct. In fact,
valid estimates of the overall
population variance of individual
samples can be derived from
composite sample measurements.
Also, identification of a 'primary'
sample is not meaningful in this
situation because the analyses were
conducted in blind random order.
Pascoe
The selection of statistical tests in Section 2 should include
mention of why the US EPA program ProUCL was not used
for determining means and other statistical metrics. The
ProUCL program includes algorithms to account for non-
detected values, with a recommended general approach that
uses the distribution of detected values to replace the non-
detects, and it can perform various statistical tests. The
ProUCL manual recommends using the replacement
procedure for non-detects if the sample size is at least six, and
for most seafood types in the CM study the number of
specimens usually exceeded this minimum value. The reliance
on use of one-half the detection limit, the full detection limit,
or in some cases a value of zero, for samples that had no
detections might not be necessary or of lesser need if ProUCL
were used.
Page 11, top partial paragraph - The sample Group ID 237
should be classified as a detect, since two of the three samples
The ProUCL statistical software is
a comprehensive statistical
software package with statistical
methods suitable for addressing a
wide range of environmental
statistical issues. The software was
developed primarily to support
analyses used evaluate the
attainment of cleanup standards for
soils and solid media, analysis of
groundwater monitoring data at
RCRA facilities and analysis of
data on chemical concentrations in
soil at CERCLA sites and exposure
data at hazardous waste sites. The
extensive ProCal capabilities were
not necessary for this study which
was, by intent, a straight forward
descriptive analysis of the fish
tissue concentration data. In this
study, upper confidence limits on
the means of the mercury levels in
fish were calculated in a straight
forward manner using simple
estimates of the population
standard deviations recovered from
the composite sample results under
the assumption of approximate
normality which is reasonable for
the composite sample results.
Authors agree.
17
-------
Responses to Charge Questions
had detected values. Referring to the final averaged value as a
non-detect seems to be an inaccurate portrayal of the intended
result, in that mercury was mostly detected.
Page 11, second full paragraph - A definition of harmonic
mean would be helpful, since, although a rarely used term, it
is becoming increasingly utilized in chemical contamination
studies. An additional mention of why it is less influenced by
outliers would also be helpful.
Table 4 - The number of composite samples for
Flounder/Fluke/Sole and for Squid under both "N.D." and
"Total" do not add up.
Table 6 - Reason should be provided in a footnote as to why
some variances are 0.0.
We should provide a definition:
The harmonic mean is the
reciprocal of the arithmetic mean
of a number of values. For n
values, KI,..., xn the harmonic
mean H = -—-—- . The
xi '" xn
harmonic mean is less influenced
by extreme values because by
definition (i.e., in mathematical
terms) it is a lower bound on the
median of a set of values. As
stated on page 11 of the report, the
harmonic mean was used to
represent the number k of fish
specimens in a composite and is a
conservative choice to represent
the number in a composite for the
purpose of estimating variance
across composite samples. This is
illustrated in the tuna example
described on page 11. The
discussion of the harmonic mean is
now on page 4 of the document
and reflects this comment
response. Some additional words
were added.
Flounder/Fluke/Sole numbers
were revised to indicate 1 non
detect and a total sample size of
14. Squid will be revised to
indicate 1 non detect and a total
sample size of 11.
The reason some variances are 0 is
that all the values in the calculation
are the same or there is only one
value.
Stern
The section on the derivation of the statistical approach is
probably more detailed and theoretical than necessary,
particularly given that the low Hg concentrations in many
species offish and the low detection limits, makes concerns
The reviewer demonstrates a
correct understanding of the use of
composite sampling and the
objectives of the study, the
straightforward nature of the
18
-------
Responses to Charge Questions
about the treatment of non-detects of little practical
significance. Nonetheless, the treatment appears reasonable
and, in the end, straightforward. The use of composite
samples provides for a reasonable estimate of the mean given
limited sampling and analytical resources. However, despite
the statistical approach that yields an estimate of the standard
deviation in the sample, composite samples necessarily
preclude accurate assessments of the extent of divergence of
individual samples within the composite from the overall
mean. This is only partially remedied by comparing the
variability among several different composites for the same
species that are ultimately combined to estimate the mean. For
some contaminants (e.g., PCBs) for which the toxicological
concern is largely with long-term average exposure, this is
generally not a problem. However, for contaminants, such as
MeHg, where short-term elevations during a critical window
of development may have toxicological significance, lack of
information on the full range of variability across individual
samples is more of a concern. While this is a shortcoming of
the statistical design of the study, it is recognized that the
specific goal of the study was to estimate mean Hg
concentrations and the variance around those mean estimates
rather than to estimate maximum concentrations or upper
percentiles of the overall distribution.
In general, the figures in the report are well presented and
informative. Figure 2 is well designed but a bit cramped and
would benefit from having one category graphic per page. The
Figure 3 presentations in Appendix A are particularly
informative.
methods employed. We disagree
with the assertion that "lack of
information on the full range of
variability across individual
samples is more of a concern" and
that "this is a shortcoming of the
statistical design of the study."
The goal of the study was to obtain
estimates of mean concentrations
which is consistent with the use of
composite sample results and the
objective of assessing possible
chronic health effects. While
obtaining individual fish
measurements was not a goal of
the study, it is possible to obtain an
estimate of the population standard
deviation from the composite data
and this was done to construct the
upper confidence limits on the
means.
This responds well to the comment
and edits to the document are not
needed.
19
-------
Responses to Charge Questions
5. Please comment on the data summary in Table 14. "Estimated Fish Servings per Week
for an Adult Female of Child-bearing age based on Means and Upper 95% Confidence
Limits on Mean Mercury Concentrations by Species". This table will draw a lot of
attention. Is the table clear and does it provide the appropriate message?
Reviewer
Comments
Response to Comments
Anderson
The table is clear and having whole meals as the metric is far
superior to the old table with fractional meals. I would place
greater emphasis on the fact that each of these only applies if
this is the only fish consumed in a given week and that these
are not to be mixed. While it is in the foot note, it is easy to
miss. It probably should be in the title of the table. This is an
issue in advisories that easily and regularly confuses the
public. I would also place this at the front of the text section
and not just at the end of the section.
However I think the message could be clearer. The upper
95% confidence interval of the mean adds little information
and is not well understood as to what it means. On the other
hand, presenting the value for two standard deviations from
the estimated population mean as a second metric would be
far more informative (as in Table 2 p 13 or figure 1 p 35).
Most people don't want to know the average fish, but what
the level is in the fish they are buying and how likely it is that
the fish they will buy is over a given value. So a measure of
the range of values is most useful. Providing the two standard
deviations from the population mean would provide the
information that going to the store and buying a fish you can
be nearly certain that it will have less mercury than this value.
That is more useful information than how robust is the
estimate of the mean. With the mean, half the fish you buy
will be above the value and of course half could be below.
While over a prolonged period the mean is the best estimate
of total dose received, it must be kept in mind that this report
and information is targeted to pregnant women and peak fetal
vulnerability may be only a short period of time. For most of
the fish in the study, these two values (upper 95% confidence
limit of the mean and two standard deviations from the
estimated population mean) are quite similar and the advice
doesn't change much. But for those fish with a high variance
it is important. It helps explain why Tuna and Swordfish are
"do not eat." And this might also help explain why blood
mercury levels were found elevated. If you have a large
family, you might be more likely to purchase a larger fish of a
species or select a larger fillet. So what you actually purchase
may not be random. The data does show that larger (or at
least longer) fish have higher concentrations within a species.
The mean minimizes that information.
The reviewer points to an
important proviso to the table, but
moving it from a footnote to the
title of the table would be
unwieldy.
We have a reasonable approach
using the confidence limits on the
mean. Anderson states the concept
that underlies our approach:
"While over a prolonged period the
mean is the best estimate of total
dose received" but is apparently
more concerned about individual
fish that could have high levels of
mercury. This was not our focus
and it is stated adequately in the
report. Do we have any
information on how harmful a
single or a few meals of fish with
high mercury may be or how this
may relate to "peak fetal
vulnerability"? Statistically we are
on more solid ground making
confidence limit statements about
the mean with these data than we
would be talking about upper
percentile estimates based on what
in many cases are rather small data
sets. Also, Anderson has confused
the mean with the median (half
above and half below) and 'what
you actually purchase may not be
random'. Table 2 shows a number
of informative statistics but does
not include values of two standard
deviations from the mean.
The upper 95% limit on the means
was used because the goal was to
estimate average consumption
levels that would represent chronic
20
-------
Responses to Charge Questions
exposures. The UCL represents a
plausible upper bound on the mean
consistent with the observed data.
The range of observed values may
be of interest but not that
informative with regard to the
content of a particular future fish
purchase.
Authors agree to pregnant women
as a sensitive subpopulation
conservative indicator and should
be a point of emphasis.
Tuna and sword fish have higher
mean levels. If you tend to eat
larger fish you will tend to intake
more Hg, the mean does not
minimize 'that information'.
Pascoe
This table is recognized as a critical summary and
presentation of the main points of the study, and as such is
clear and appropriate.
No response needed.
Stern
In terms of its presentation, Table 14 is clear and easily
interpretable. However, there is an underlying potential for
significant uncertainty in the interpretation and application of
the results of Table 14 relative to the FDA/EPA guidance to
which it implicitly relates. The FDA/EPA advisory language
for Hg in commercial fish states that, with a few specific
exceptions, 12 oz., equal to two meals, offish should be eaten
per week. Thus, the assumed nominal average serving size is
6 oz. However, due to loss of moisture content with cooking,
a 6 oz. cooked serving offish is approximately equivalent to
8 oz. of purchased raw fish. Since the language in the
FDA/EPA advisory does not specify as-purchased, or as-
cooked, the actual intended serving size is unclear. Table 14
is based on 8 oz. wet weight (i.e., as-purchased). Presumably,
the intended guidance to be derived from Table 14 relates to
as-purchased weight. This is appropriate since stores weigh
fish on purchase, but consumers generally do not weigh fish
once it is cooked. However, this can lead to confusion with
respect to comparison of Table 14 to the FDA/EPA guidance.
Another issue with Table 14 is that (as stated on pg. 37,
paragraph 2) the number of meals per week so as not to
exceed the RfD is presented relative to two different estimates
of the Hg concentration by species - the mean and the mean +
2 population-sd. However, Table 14 presents the latter
category as the 95% upper confidence limit (UCL) on the
mean. The 95% UCL of the mean is equal to the standard
error of the mean times the t-statistic corresponding to 95% of
the t-distribution. Thus, the mean + 2 sd is approximately
equivalent to the 98th percentile of the population distribution,
To clarify a point raised by the
reviewer, the 8 oz. meal relates to
as purchased (uncooked) or wet
weight to be consistent with the H
analysis which was performed on
raw specimens.
Address the 95 % UCL vs mean +
2 SD issue.
21
-------
Responses to Charge Questions
and not the 95* UCL on the mean. Given the relatively small
number of observations, this is not likely to make a large
practical difference, but this should be corrected.
Also, as discussed above, the implications and intended use of
Table 14 as presented in the report are ambiguous. Strictly
speaking, Table 14 applies only to consumers who get their
commercial fish from the NYC wholesale market. Thus,
guidance to be derived from Table 14 would specifically
apply only to those consumers. Further, apparent
contradictions between the FDA/EPA guidance and Table 14
would not necessarily imply contradictions if the nationwide
wholesale fish supply were similarly examined. The nominal
specificity of Table 14 should be clearly stated and the
implications or lack of implications for the application of the
FDA/EPA advisory nationally should also be clearly stated.
As noted previously, the results of
this study were not intended to be
generalizable to the nation's
seafood supply and the
recommendations for meal
frequency of different seafood
species is not intended to be a
referendum on FDA regulations
and/or guidance.
6. Were the analytic methods for obtaining mercury and PCB concentrations in fish tissue
appropriate for this study?
Reviewer
Comments
Response to Comments
Anderson
The laboratory methods used were standard and those widely
used. Doing congener specific PCB analyses was informative
although no analyses were included looking at whether
congener patterns were different by species etc. The
variability analyses demonstrated that the analytic procedures
were applied consistently and the laboratories used excellent
QA/QC practices.
The PCB analysis was ancillary to
the primary goal of the study;
therefore, analysis of PCB congener
patterns across species is beyond the
study's scope.
Pascoe
The analytical methods were appropriate for total mercury
and PCBs. PCBs were appropriately analyzed as congeners.
For mercury, the detection limits were sufficiently low
enough for adequate detection in most seafood tissues.
No response needed.
Stern
The analytical method for Hg determination in fish tissue
appears to be the standard cold vapor AA method. However,
there is no mention of standard reference materials used in
calibration or QA. I assume that these were, in fact, used, but
this should be stated.
With respect to the PCB analyses, the analytical method
appears to be relatively standard and the approach of adding
total PCB congener concentrations to obtain a composite
measure of total PCBs is appropriate for the purpose of
comparison to the FDA tolerance limit. However, it should
Text was added to reflect the
reviewers comment re: co-planer
(dioxin-like) PCB congeners.
22
-------
Responses to Charge Questions
be noted that certain PCB congeners have toxicity that relates
to TCDD-TEQ derived measures in addition to their toxicity
as measured by total PCBs.
The significance and utility of the lipid adjusted metric of
PCB concentration should be addressed. Notwithstanding
that PCB concentrations are often reported relative to lipid
adjustment, from a consumption/exposure standpoint, why
would lipid adjustment be useful?
23
-------
Responses to Charge Questions
7. Given that the fish specimens were obtained from commercial venders and some
specimens were not whole fish, the EPA implemented DNA analysis for verification of
species. Please comment on this approach and how the information is presented in the
appendix of the report.
Reviewer
Comments
Response to Comments
Anderson
The use of DNA analyses to confirm fish identities was quite
innovative and informative. Yet the only mention of the
DNA results is in the executive summary. Appendix B,
while very informative, does not interpret the results. The
conclusions reached and briefly mentioned in the executive
summary need to be enhanced and maybe made a section in
the report. Although not directly stated, it does appear that
fish are not being mislabeled in a manner that would lead to
fish with higher mercury levels being sold as a lower
contaminated species or via versa. That is important
consumer information. However, it would be useful to see a
discussion as the 27 samples that were reanalyzed and 5
were changed and found to agree with the morphologically
derived result. What is of interest is what of the 22 others
that apparently remained different than the description of the
fish. While less than 10% of the samples analyzed, it would
be good to know how these differed. Are these really
different or is it a method issue?
DNA barcoding section was not
moved from the appendices to the
body of the report.
Pascoe
The use of barcoding for this study was a fairly novel use of
the technique and appropriate to the study. Typical fish
sampling studies would rely on fish taxonomists to identify
specimens either during field collection or in the laboratory
prior to tissue dissection. As mentioned above, the utility of
the barcoding exercise could be emphasized more in the
report, that difference between taxonomic identification and
the barcoding also resulted in EPA's re-examination of the
photos of some of the collected specimens to determine
whether they had been mis-identified.
Added that Moses Chang, an
ichthyologist, was at the
commercial fish market during
sample acquisition.
Stern
There is clearly a potential for misidentification and
mislabeling offish species at the commercial level. Thus,
DNA barcoding was an appropriate approach for positive
species identification. It should be noted, however, that the
FDA database to which the barcode-identified species are
compared is not based on similar species identification.
Furthermore, consumers purchase fish based on the vendor's
identification that is, likewise, generally not based on DNA
barcoding. Thus, while the DNA barcoding carried out in
this study is a useful first step in putting fish consumption
advisories on a standardized footing, it does not, practically
speaking, reduce the uncertainty for comparison with the
No response needed.
24
-------
Responses to Charge Questions
FDA database or for the consumer in purchasing.
The description of the barcoding procedure and the reporting
of the results of the barcoding in the report are inconsistent
with respect to the level of technical detail. Some of the
information on the barcoding procedure and interpretation
(e.g., the CBOL criteria, the description of the primers and
sequence overlap, the description of the result qualifier
scores) are quite technical and largely inaccessible to even
scientific readers who are not previously acquainted with the
terminology and procedures. Other parts (e.g., explanation of
DNA structure and base-pairing) are much more
straightforward and apparently aimed at a different audience.
In the end, all of the method description and data
presentation does not lead to a clear discussion of the
barcoding findings and their implications.
25
-------
Appendix A: Individual Reviewer Comments
A-l
-------
-------
COMMENTS SUBMITTED BY
Henry A. Anderson, M.D.
Private Consultant
Madison, WI 53704
608-241-1227
Email: anderha@sbcglobal.net
A-3
-------
-------
Henry A. Anderson, M.D.
Fish Tissue Analysis for Mercury and PCBs from a New York City
Commercial Fish/Seafood Market
Contract No. EP-C-07-024
Task Order No. 114
June 3 0,2011
External Letter Peer Review of EPA's Draft Report
CHARGE QUESTIONS
General note: There are a few typos and what appear to be mislabels in the document. The document
needs careful editing to correct and find all these gremlins. For instance on page 34 a "Figure 10 in
Appendix A " is mentioned, but there is no Figure 10 in appendix A. I believe this is referring to what is
labeled Figure 3. It appears that tables and figures are numbered consecutively from the beginning
through appendix A, then B appendix starts numbering again as does C. That is confusing.
1. Please comment on the organization and clarity of the report.
The organization of the report is appropriate and easy to follow. There are a number of approaches that
would help with clarity. The report is quite long and presents many tables and figures which are very
useful. This makes the executive summary become particularly important. Right now the executive
summary is pretty cursory and the key findings of the report need to be highlighted in some kind of order.
Having them as bullets would improve the clarity. It might also help if the key findings were highlighted
in each section of the report at the front. These could be bullets but listing them and then having the text
explain and elaborate would help. For instance, there is a lot of interest in farmed vs wild fish. This is
mentioned and information provided in the text and there is not much to say since only three species were
farm raised, and only one of those a fin fish. But it might be mentioned that no mercury was detected in
the wild Atlantic salmon but found in half of the farmed salmon, even though the levels were very low.
Farm vs wild is only included as part of "water body" discussion. If the authors go through each section
and select what they view as the key findings, list them in the sections and then carry them into the
executive summary that would help the reader focus on key findings.
The impetus for the study was the New York City blood mercury report. This study and its findings
should probably be summarized a bit better and the fish consumption rates mentioned as well as the
mercury prevalence's in the general population and other associated factors. The question that this report
does not address is whether these fish monitoring data presented can explain the blood mercury
distributions seen? It would help if somehow the authors could translate the various "action levels"
A-5
-------
Henry A. Anderson, M.D.
presented into what are their blood mercury level equivalents. Or at least discuss how the meal rates
reported in Table 14a might translate into blood mercury.
What does the NY 5 ug/L blood translate into as fish contaminant level and meals per week? Will the
Table 14a result in blood mercury less than the NY 5 ug/L value? It is somewhat counter intuitive that
none of the composites exceed the FDA action level, and very few the EPA Rfd based level, yet 72% of
the NYC Chinese exceed the NY reportable level of 5 ug/L mercury in blood. If these "action levels" are
to protect everyone at unlimited fish consumption levels, then these "action levels" don't seem to be
protective as far as the NY reportable level is concerned. Some discussion of the basis for the NY
reportable level is needed and how it relates to the report's findings. In the executive summary it is
mentioned that the NYC fish samples appear to be somewhat lower than the FDA comparisons. That
triggers the question of then why do NYC residents have blood mercury levels three times those of the
rest of the country? I would suggest that the testing samples are qualitatively similar, but doing a
statistical comparison is problematic because of the different sampling structure.
In the introduction it is mentioned that the "action levels" used will be discussed further in the report. But
all that is included is a table and references. There is no discussion of how they were derived, why they
are different and how they are intended to be used. This is a bit of apples vs oranges. What is the level of
fish consumption that each assumes? The challenge in understanding this report is how to arrive at a
"dose" by navigating between fish tissue concentration, fish consumption rates over a selected time
period and the blood mercury levels seen in the NYC study. It appears simple but is complex.
2. Is the data adequate for meeting the objectives of this study?
It is unclear whether the data meets the object of measuring mercury in commonly consumed seafood
species in New York. Certainly the data represents the fresh seafood consumed in the city. What is not
addressed is what percentage of seafood consumed is consumed fresh. If there is any data on total seafood
from the city or other surveys, reporting that would help place this data in perspective. The data is
certainly important and very useful. The data and methods descriptions are very good and comprehensive.
It is easy to understand what was done and mostly why. It should be helpful to the NYC "Eat Fish,
Choose Wisely" project.
The PCB information, although an ancillary activity is very useful and provides a great deal of
information that augments the paucity of commercial fish PCB data. Most of the PCB mention is
relegated to the appendix C except for brief mention in the executive summary. The critical statement in
A-6
-------
Henry A. Anderson, M.D.
the executive summary is that the PCBs appear within the FDA tolerance level. For mercury the authors
selected several "action levels" to compare to not just the FDA tolerance. Many states, as well as the EPA
have "action levels" for PCBs which are, as with mercury, considerably lower than what FDA uses.
While the PCBs are all quite low, including the EPA RfD or some of the states' or the Great Lakes
Protocol value would be informative. Several states have catfish on their advisories. The highest catfish
with PCB appears to be a wild caught lake fish. So for comparative purposes, having the low to high
"action levels" used by various entities for PCB would be appropriate.
3. Is the selection of the fish species adequate for this study?
The selection offish species is adequate for the general NYC population who purchase fresh fish. The
sample sizes while not large are sufficient. While analyses of single fish would have been preferred
simply because that is what is most descriptive of consumer activity, the use of composites is acceptable.
This does make it difficult to directly compare results from other studies. But, the statistical analyses
provided show that the findings are consistent with what has been found in other studies. It would be
helpful to have a bit more explanation of how the use of composites rather than the same number of
samples but of individual fish strengthens the findings. No major commercial species appears to be
missing.
Given that the report introduction and background emphasizes the blood mercury levels in the NYC
Chinese, what is missing is any sampling of the imported frozen, canned and dried fish that are sold in
Chinese/oriental groceries. Some mention of any published studies showing that these fish are similar in
contaminant concentration than other fish would help. It is unclear how much of the fish in Chinese diets
come from the species sampled in this study. The lack of ethnic specific commercial fish should at least
be discussed and might help explain the biomonitoring observations rather than just that for this
population frequency of consumption probably accounts for the differences. At least some discussion or
at least the recognition of the possible impact of ethnically preferred fish needs to be mentioned. As
mentioned in response to #1, it would be good to include the biomonitoring blood mercury data for the
general NYC population and not just the Chinese. This study is not ethnic focused.
4. Is the use and presentation of the descriptive statistics appropriate?
The use of descriptive statistics is appropriate and when reported are presented with appropriate caveats
about sample size etc and explanations of why certain procedures were or were not used. This was very
helpful.
A-7
-------
Henry A. Anderson, M.D.
The description of the analyses of variability is quite good and comprehensive. It is probably too detailed
for most readers and all the formulae are quite intimidating. These details might be better included as an
appendix and just the results and general methods discussed in the main text. What is also needed is a
discussion of the a priori acceptable level of variability or what variability is scientifically acceptable. A
clear set of concluding statements are needed. These need to indicate that the laboratory analyses and
sample processing methods were assessed for consistency and reliability and found to meet accepted
scientific standards. Actually the minimal variability found is quite impressive and shows the rigorous
quality control that was utilized.
I don't think it is appropriate to add the replicates and duplicates to the primary sample and average them
and then assign that value to the sample. This really messes up the variability since they are no longer
individual composite results but means. I doubt it makes much practical difference, but I would prefer that
the primary sample result be the only result used in the overall analyses. Then all the composite sample
results are truly comparable. The replicates and duplicates should be used only for what they were
intended to be used for - assessing variability. The report quite effectively presents the impact of various
analytic decisions such as how to handle the non-detects - showing that the differences in methods make
little impact on conclusions. The same "with and without" approach could be done for the
duplicate/replicate averaging impact. Rather than redo all the tables without the replicate/duplicate data
included, you could see if it makes a difference and if it does not, simply state that to be the case. Then
explain why you believe having some values as means and others not is more robust and statistically
appropriate and that is why you chose to "use all the data". But whether it is appropriate and allows use of
the statistical procedures used in other tables, I will leave to a statistician to decide.
5. Please comment on the data summary in Table 14, "Estimated Fish Servings per Week for
an Adult Female of Child-bearing age based on Means and Upper 95% Confidence Limits
on Mean Mercury Concentrations by Species". This table will draw a lot of attention. Is the
table clear and does it provide the appropriate message?
The table is clear and having whole meals as the metric is far superior to the old table with fractional
meals. I would place greater emphasis on the fact that each of these only applies if this is the only fish
consumed in a given week and that these are not to be mixed. While it is in the foot note, it is easy to
miss. It probably should be in the title of the table. This is an issue in advisories that easily and regularly
confuses the public. I would also place this at the front of the text section and not just at the end of the
section.
A-8
-------
Henry A. Anderson, M.D.
However I think the message could be clearer. The upper 95% confidence interval of the mean adds little
information and is not well understood as to what it means. On the other hand, presenting the value for
two standard deviations from the estimated population mean as a second metric would be far more
informative (as in Table 2 p 13 or figure 1 p 35). Most people don't want to know the average fish, but
what the level is in the fish they are buying and how likely it is that the fish they will buy is over a given
value. So a measure of the range of values is most useful. Providing the two standard deviations from the
population mean would provide the information that going to the store and buying a fish you can be
nearly certain that it will have less mercury than this value. That is more useful information than how
robust is the estimate of the mean. With the mean, half the fish you buy will be above the value and of
course half could be below. While over a prolonged period the mean is the best estimate of total dose
received, it must be kept in mind that this report and information is targeted to pregnant women and peak
fetal vulnerability may be only a short period of time. For most of the fish in the study, these two values
(upper 95% confidence limit of the mean and two standard deviations from the estimated population
mean) are quite similar and the advice doesn't change much. But for those fish with a high variance it is
important. It helps explain why Tuna and Swordfish are "do not eat." And this might also help explain
why blood mercury levels were found elevated. If you have a large family, you might be more likely to
purchase a larger fish of a species or select a larger fillet. So what you actually purchase may not be
random. The data does show that larger (or at least longer) fish have higher concentrations within a
species. The mean minimizes that information.
6. Were the analytic methods for obtaining mercury and PCB concentrations in fish tissue
appropriate for this study?
The laboratory methods used were standard and those widely used. Doing congener specific PCB
analyses was informative although no analyses were included looking at whether congener patterns were
different by species etc. The variability analyses demonstrated that the analytic procedures were applied
consistently and the laboratories used excellent QA/QC practices.
7. Given that the fish specimens were obtained from commercial venders and some specimens
were not whole fish, the EPA implemented DNA analysis for verification of species. Please
comment on this approach and how the information is presented in the appendix of the
report.
The use of DNA analyses to confirm fish identities was quite innovative and informative. Yet the only
mention of the DNA results is in the executive summary. Appendix B, while very informative, does not
interpret the results. The conclusions reached and briefly mentioned in the executive summary need to be
A-9
-------
Henry A. Anderson, M.D.
enhanced and maybe made a section in the report. Although not directly stated, it does appear that fish are
not being mislabeled in a manner that would lead to fish with higher mercury levels being sold as a lower
contaminated species or via versa. That is important consumer information. However, it would be useful
to see a discussion as the 27 samples that were reanalyzed and 5 were changed and found to agree with
the morphologically derived result. What is of interest is what of the 22 others that apparently remained
different than the description of the fish. While less than 10% of the samples analyzed, it would be good
to know how these differed. Are these really different or is it a method issue?
A-10
-------
COMMENTS SUBMITTED BY
Gary A. Pascoe, Ph.D., DABT
Principal, Pascoe Environmental Consulting
Port Townsend, WA 98368
360-385-9977
Email: gpascoe@plympus.net
A-ll
-------
-------
Gary A. Pascoe, Ph.D., DABT
External Peer Review
Contract No. EP-C-07-024
Task Order No. 114
June 3 0,2011
External Letter Peer Review of EPA's Draft Report,
Fish Tissue Analysis for Mercury and PCBs from a New York City
Commercial Fish/Seafood Market
1. Please comment on the organization and clarity of the report.
The report is well organized and clearly written, with a few exceptions noted below; nonetheless, it
should be easily understandable by health professionals and the interested public. The following are
editorial comments that either need addressing or may help with clarity. These comments focus on the
Executive Summary, which should be both comprehensive yet simplified enough with explanations of
technical issues in order to be sufficiently understandable to someone unfamiliar with those issues.
Page 1, Executive Summary, first line - Because some states use the same or similar regulatory agency
name, EPA should be referred to as the U.S. Environmental Protection Agency upon first use in the
Executive Summary and the main body of the report.
Page 1, Executive Summary, first paragraph - The goal of the study to support the NYC public health
message should be clarified with more detail. The development of a large seafood mercury database to
provide support for the City's message on public health is highly admirable, and the readers could be
more informed of how the data were actually used in formulating the health messages, and what those
messages are. The final tables in the report provide recommendations on meals per month that sensitive
members of the population (e.g., women of child bearing age) can consume of different types offish from
the Commercial Market, and the development of these recommendations using the mercury seafood data
that the study collected should be more strongly linked to the NYC public health messages.
Page 1, Executive Summary, second paragraph - The derivation and meaning of the term "reportable
level" is not explained, and may have different meanings to a chemist and a health professional or
someone not versed in those fields. The term should be clarified as to whether it is a health-based
criterion or a departmental advisory, or an enforceable standard, or whether it refers to concentrations that
should be reported to a regulatory agency, and if so the basis for the requirement and the concentration of
mercury should be noted.
A-13
-------
Gary A. Pascoe, Ph.D., DABT
Page 1, Executive Summary, third paragraph, fourth sentence - A comma appears to be missing.
Page 2, Executive Summary, third paragraph - The first sentence appears to state that the report estimates
the amount of seafood adult women of child bearing age consume, from which the intake of mercury is
estimated. This is an incorrect description of the process that is actually used in the report, in that the
amount of seafood consumed by adult women of child bearing age is not itself estimated, but instead what
is estimated is the amount of seafood consumption that corresponds to a level of concern for exposure to
mercury. If the report intended to estimate the amount of seafood that a population or subgroup of a
population consumes, an entirely different type of consumption study would be needed. The amount of
seafood consumption considered to correspond to a level of concern is independent of actual seafood
consumption rates. This distinction is critical primarily because some subpopulations will consume more
seafood than others, which is noted at the end of the report, and descriptions of seafood consumptions by
subpopulations need to be sensitive to such cultural differences. The fourth sentence of this paragraph
also alludes to a determination of the amount of seafood that a subpopulation consumes - the permissible
daily intake of methyl mercury was not actually compared with the amount of methyl mercury intake
from fish ingestion, but rather was used to identify the permissible amount of ingestion of fish that
contain varying levels of methyl mercury.
Page 2, Executive Summary, fourth paragraph and Section 6 - The report should attempt to provide some
explanation of why the NYC CM data were so much lower than the FDA data on mercury in fish tissue, if
there is a known or suspected reason.
Page 3, Executive Summary, first full paragraph - Additional summary of the rationale for the use of the
barcoding would be appropriate here. A rationale is alluded to at the beginning of Appendix B which
discusses the problems of taxonomic identification of dead fish. It should be emphasized in the Executive
Summary that the barcoding was considered useful because the sources of the fish specimens consisted of
different commercial fishers with uncertain quality of handling of the specimens, and the sources were not
from a scientific survey and collection from the water bodies from which NYC fish are typically caught.
Additionally, there was concern that the handling of the fish specimens by the commercial fishers could
result in enough damage to the fish exterior to make it difficult to identify at a species level, as well as the
reasons stated in Appendix B. In addition, more summary of the utility of the barcoding to the study
should be added to the end of the Executive Summary. Specifically, the first full paragraph of Page 113
mentions that disparities between taxonomic identification and barcoding also resulted in re-examination
of the photos of collected specimens by EPA to determine whether they had been mis-identified.
A-14
-------
Gary A. Pascoe, Ph.D., DABT
Page 31, Section 5.2, third paragraph - This text discusses the comparison of total mercury concentrations
in fish with the EPA action level for methyl mercury. The reason given as to why EPA considers it
acceptable to compare total mercury concentrations with the action level for methyl mercury is not
entirely correct. EPA (2009) does not state that the comparison is acceptable because the molecular
weight of mercury is close to that of methyl mercury, but because most if not sometimes all of the total
mercury in fish tissue is actually methyl mercury, as stated in the text of the subsequent paragraph (fourth
paragraph of Section 5.2). The text of the third paragraph needs to be changed to reflect the actual
reasoning contained in EPA (2009) for the comparison.
Page 34 - The reference to Figure 10 in Appendix A should be Figure 3.
Page 37 - The reference to Table D-la appears to refer to Table 14.
Page 43, Section 6 has a formatting problem.
Page 54 - Two references for EPA 2009 are provided; however, the second reference of EPA (2009)
Guidance for Implementing the January 2001 Methylmercury Water Quality Criterion has been
superseded by an updated document: USEPA. 2010. Guidance for Implementing the January 2001
Methylmercury Water Quality Criterion. April. EPA 823-R-10-001.
http://water.epa.gov/scitech/swguidance/standards/criteria/aqlife/pollutants/methylmercurv/upload/mercur
v2010.pdf.
2. Is the data adequate for meeting the objectives of this study?
The data collected by NYC and EPA are fully adequate for meeting the primary objective of the study.
Based on the objective of analyzing mercury in the different seafood types that are consumed by people in
NYC, the report sufficiently characterized mercury in a large suite of seafood types, covering multiple
trophic levels and multiple phylogenetic levels. The study collected sufficient numbers of samples to
enable statistical analyses among the different categories of seafood for total mercury concentrations, and
among species within some of the categories.
In Appendix C, the data collected on PCB concentrations in fish specimens were insufficient to perform a
similar level of statistical analysis as was performed for mercury, but the data collection was not a
primary objective of the study and the data still provide useful information on levels of PCBs that the
general fish consuming public might be exposed to. The discussion on page 122 about the adequacy of the
A-15
-------
Gary A. Pascoe, Ph.D., DABT
detection limits could be expanded to include comparison of total PCBs based on assuming non-detected
congener concentrations at the detection limit rather than as zero, which would appear to still result in the
concentration of total PCBs based on congeners to be below the FDA action level.
The report mentions the availability of blood levels of mercury for NYC residents. If studies have been
performed that link those blood levels to patterns of seafood consumption and types offish, particularly
for ethnic or cultural subpopulations, some discussion should be considered in the report that would more
fully link those studies with the objectives of the CM study.
3. Is the selection of the fish species adequate for this study?
The species selected for the study address multiple types of seafood that are typically consumed by the
targeted audience, with the variety ranging from shellfish to squid to various sizes and types of finfish.
Since the objective of the study focused on the types of seafood that consumers would typically eat from a
commercial market, the variety in the types of seafood that were collected from the market was
appropriate and adequate. Because the study was able to analyze fish and shellfish from multiple trophic
and phylogenetic levels, it exceeded the needs of the study with regards to variety in seafood types.
4. Is the use and presentation of the descriptive statistics appropriate?
The selection of statistical tests in Section 2 should include mention of why the US EPA program
ProUCL was not used for determining means and other statistical metrics. The ProUCL program includes
algorithms to account for non-detected values, with a recommended general approach that uses the
distribution of detected values to replace the non-detects, and it can perform various statistical tests. The
ProUCL manual recommends using the replacement procedure for non-detects if the sample size is at
least six, and for most seafood types in the CM study the number of specimens usually exceeded this
minimum value. The reliance on use of one-half the detection limit, the full detection limit, or in some
cases a value of zero, for samples that had no detections might not be necessary or of lesser need if
ProUCL were used.
Page 11, top partial paragraph - The sample Group ID 237 should be classified as a detect, since two of
the three samples had detected values. Referring to the final averaged value as a non-detect seems to be
an inaccurate portrayal of the intended result, in that mercury was mostly detected.
A-16
-------
Gary A. Pascoe, Ph.D., DABT
Page 11, second full paragraph - A definition of harmonic mean would be helpful, since, although a
rarely used term, it is becoming increasingly utilized in chemical contamination studies. An additional
mention of why it is less influenced by outliers would also be helpful.
Table 4 - The number of composite samples for Flounder/Fluke/Sole and for Squid under both "N.D."
and "Total" do not add up.
Table 6 - Reason should be provided in a footnote as to why some variances are 0.0
5. Please comment on the data summary in Table 14, "Estimated Fish Servings per Week for
an Adult Female of Child-bearing age based on Means and Upper 95% Confidence Limits
on Mean Mercury Concentrations by Species". This table will draw a lot of attention. Is the
table clear and does it provide the appropriate message?
This table is recognized as a critical summary and presentation of the main points of the study, and as
such is clear and appropriate.
6. Were the analytic methods for obtaining mercury and PCB concentrations in fish tissue
appropriate for this study?
The analytical methods were appropriate for total mercury and PCBs. PCBs were appropriately analyzed
as congeners. For mercury, the detection limits were sufficiently low enough for adequate detection in
most seafood tissues.
7. Given that the fish specimens were obtained from commercial venders and some specimens
were not whole fish, the EPA implemented DNA analysis for verification of species. Please
comment on this approach and how the information is presented in the appendix of the
report.
The use of barcoding for this study was a fairly novel use of the technique and appropriate to the study.
Typical fish sampling studies would rely on fish taxonomists to identify specimens either during field
collection or in the laboratory prior to tissue dissection. As mentioned above, the utility of the barcoding
exercise could be emphasized more in the report, that difference between taxonomic identification and the
barcoding also resulted in EPA's re-examination of the photos of some of the collected specimens to
determine whether thev had been mis-identified.
A-17
-------
-------
COMMENTS SUBMITTED BY
Alan H. Stern, Dr.P.H., D.A.B.T.
Independent Consultant
Metchen, NJ 08840
609-633-2374
Email: ahsternl@verizon.net
A-19
-------
Review of Fish Tissue Analysis for Mercury and PCBs from a New York City
Commercial Fish/Seafood Market
Alan H. Stern, Dr.P.H.
1. Please comment on the organization and clarity of the report.
In general, the report is well organized and clearly written. There are some specific points where the text
is unclear. These are noted in my specific comments. The one significant exception to this, as discussed in
detail below, is that the full objectives of the report as evinced by the analyses in the report, itself, are not
clearly stated. Notwithstanding the overall logical organization and general clarity of the text, however,
the report seems to reflect an inconsistent level of technical detail. Some sections are technically quite
detailed and appear not to be intended to be generally accessible to the lay readier such as, for example,
Section 4, "Analysis of Measurement Variability" and Appendix B, the explanation DNA "barcode"
analysis. While Section 5, "Comparison of CM Data to Risk Metrics" is much more accessible. Some
thought could be given to the intended audience for this report.
The report compares the NYC commercial market Hg concentrations by species to the FDA database of
Hg concentration by species and gives the implications for consumption frequency advisories relative to
the EPA RfD. Further, the report finds that, in general, the concentrations from the NYC commercial
market samples are significantly lower on a species-by-species basis. The obvious next logical step is to
compare the current FDA/EPA advisory specifics to the meal frequencies derived in the report (Table 14).
This is not done and is conspicuous by its absence.
2. Is the data adequate for meeting the objectives of this study?
The specific objectives of the study do not appear to be stated in the report. This is a shortcoming that
should be rectified. The first paragraph of Section 1 states that study was ".. .undertaken to measure
mercury (Hg) concentration in composite samples from seafood species most commonly consumed by
New York City residents as represented by specimens obtained from a commercial market." Based on the
sections of the report that compare the NY City market Hg concentrations to those from the same species
in the FDA database and based on the calculation of the meal frequency by species that would exceed the
RfD, it is clear that the objectives of the report go beyond simply measuring Hg concentration in common
commercial fish. The apparent objective of the study is to assess the currency and accuracy of the existing
FDA/EPA mercury advisory for commercial fish and to recommend possible adjustments based on the
more up-to-date NY City commercial market data. To the extent that the objectives are focused on the
A-20
-------
FDA database and on the FDA/EPA advice relative to the NYC wholesale market per se, the study
appears to meet those implied objectives both in terms of sampling and statistical analysis.
However, given the absence of clear statement of objectives, it is not clear to what extent EPA intends the
data from the NYC commercial market to be representative of commercial fish Hg data nationwide.
Section 1 does state the goal of the study was to measure Hg levels in fish commonly consumed by NYC
residents. Nonetheless, the study clearly has broader implications. The NYC commercial fish market is a
major intake and distribution point for commercial fish in the eastern U.S. and the NYC wholesale market
receives fish caught in global waters. Nonetheless, there are other wholesale commercial fish markets in
the eastern U.S. and there are other wholesale markets that serve other parts of the U.S. Thus, it is unclear
to what extent EPA intends to generalize the findings of this study to the U.S. as a whole. A clear
statement of the objectives of this study should include a discussion of the intent of its focus on the NYC
market and the extent to which EPA believes that data from the NYC market can and should be used to
generalize to commercial fish nationwide.
3. Is the selection of the fish species adequate for this study?
As above, it is not possible to tell the extent to which the selection offish species is adequate without
knowing the overall objective of the report and more specifically, without knowing the extent to which
EPA intends to generalize from the data collected from the NYC commercial market. The stated intent of
the collection process was to obtain a representative sample of the fish most commonly consumed by
NYC residents. The report briefly cites several databases (NMFS fish landings data; "Fish availability in
supermarkets... in New Jersey" (Burger et al., 2004, but not specifically identified as such in the report);
and the "Seafood Products Matrix" (not further defined)) as the basis for the selection offish obtained
from the NYC market. Given the level of detail in other parts of the report, it is surprising that no
information was provided as to how these databases were actually used to determine the market sampling
strategy. This information should be included in the report. The list of fish included in the study appears
to be appropriate and does, in my subjective assessment, include all or most fish commonly found in
NYC markets. Nonetheless, given the nature of the study, documentation of the process seems essential.
4. Is the use and presentation of the descriptive statistics appropriate?
The section on the derivation of the statistical approach is probably more detailed and theoretical than
necessary, particularly given that the low Hg concentrations in many species offish and the low detection
limits, makes concerns about the treatment of non-detects of little practical significance. Nonetheless, the
A-21
-------
treatment appears reasonable and, in the end, straightforward. The use of composite samples provides for
a reasonable estimate of the mean given limited sampling and analytical resources. However, despite the
statistical approach that yields an estimate of the standard deviation in the sample, composite samples
necessarily preclude accurate assessments of the extent of divergence of individual samples within the
composite from the overall mean. This is only partially remedied by comparing the variability among
several different composites for the same species that are ultimately combined to estimate the mean. For
some contaminants (e.g., PCBs) for which the toxicological concern is largely with long-term average
exposure, this is generally not a problem. However, for contaminants, such as MeHg, where short-term
elevations during a critical window of development may have toxicological significance, lack of
information on the full range of variability across individual samples is more of a concern. While this is a
shortcoming of the statistical design of the study, it is recognized that the specific goal of the study was to
estimate mean Hg concentrations and the variance around those mean estimates rather than to estimate
maximum concentrations or upper percentiles of the overall distribution.
In general, the figures in the report are well presented and informative. Figure 2 is well designed but a bit
cramped and would benefit from having one category graphic per page. The Figure 3 presentations in
Appendix A are particularly informative.
5. Please comment on the data summary in Table 14, "Estimated Fish Servings per Week for
an Adult Female of Child-bearing age based on Means and Upper 95% Confidence Limits
on Mean Mercury Concentrations by Species". This table will draw a lot of attention. Is the
table clear and does it provide the appropriate message?
In terms of its presentation, Table 14 is clear and easily interpretable. However, there is an underlying
potential for significant uncertainty in the interpretation and application of the results of Table 14 relative
to the FDA/EPA guidance to which it implicitly relates. The FDA/EPA advisory language for Hg in
commercial fish states that, with a few specific exceptions, 12 oz., equal to two meals, offish should be
eaten per week. Thus, the assumed nominal average serving size is 6 oz. However, due to loss of moisture
content with cooking, a 6 oz. cooked serving offish is approximately equivalent to 8 oz. of purchased raw
fish. Since the language in the FDA/EPA advisory does not specify as-purchased, or as-cooked, the
actual intended serving size is unclear. Table 14 is based on 8 oz. wet weight (i.e., as-purchased).
Presumably, the intended guidance to be derived from Table 14 relates to as-purchased weight. This is
appropriate since stores weigh fish on purchase, but consumers generally do not weigh fish once it is
cooked. However, this can lead to confusion with respect to comparison of Table 14 to the FDA/EPA
guidance.
A-22
-------
Another issue with Table 14 is that (as stated on pg. 37, paragraph 2) the number of meals per week so as
not to exceed the RfD is presented relative to two different estimates of the Hg concentration by species -
the mean and the mean + 2 population-sd. However, Table 14 presents the latter category as the 95%
upper confidence limit (UCL) on the mean. The 95% UCL of the mean is equal to the standard error of
the mean times the t-statistic corresponding to 95% of the t-distribution. Thus, the mean + 2 sd is
approximately equivalent to the 98th percentile of the population distribution, and not the 95th UCL on the
mean. Given the relatively small number of observations, this is not likely to make a large practical
difference, but this should be corrected.
Also, as discussed above, the implications and intended use of Table 14 as presented in the report are
ambiguous. Strictly speaking, Table 14 applies only to consumers who get their commercial fish from the
NYC wholesale market. Thus, guidance to be derived from Table 14 would specifically apply only to
those consumers. Further, apparent contradictions between the FDA/EPA guidance and Table 14 would
not necessarily imply contradictions if the nationwide wholesale fish supply were similarly examined.
The nominal specificity of Table 14 should be clearly stated and the implications or lack of implications
for the application of the FDA/EPA advisory nationally should also be clearly stated.
6. Were the analytic methods for obtaining mercury and PCB concentrations in fish tissue
appropriate for this study?
The analytical method for Hg determination in fish tissue appears to be the standard cold vapor AA
method. However, there is no mention of standard reference materials used in calibration or QA. I assume
that these were, in fact, used, but this should be stated.
With respect to the PCB analyses, the analytical method appears to be relatively standard and the
approach of adding total PCB congener concentrations to obtain a composite measure of total PCBs is
appropriate for the purpose of comparison to the FDA tolerance limit. However, it should be noted that
certain PCB congeners have toxicity that relates to TCDD-TEQ derived measures in addition to their
toxicity as measured by total PCBs.
The significance and utility of the lipid adjusted metric of PCB concentration should be addressed.
Notwithstanding that PCB concentrations are often reported relative to lipid adjustment, from a
consumption/exposure standpoint, why would lipid adjustment be useful?
A-23
-------
7. Given that the fish specimens were obtained from commercial venders and some specimens
were not whole fish, the EPA implemented DNA analysis for verification of species. Please
comment on this approach and how the information is presented in the appendix of the
report.
There is clearly a potential for misidentification and mislabeling offish species at the commercial level.
Thus, DNA barcoding was an appropriate approach for positive species identification. It should be noted,
however, that the FDA database to which the barcode-identified species are compared is not based on
similar species identification. Furthermore, consumers purchase fish based on the vendor's identification
that is, likewise, generally not based on DNA barcoding. Thus, while the DNA barcoding carried out in
this study is a useful first step in putting fish consumption advisories on a standardized footing, it does
not, practically speaking, reduce the uncertainty for comparison with the FDA database or for the
consumer in purchasing.
The description of the barcoding procedure and the reporting of the results of the barcoding in the report
are inconsistent with respect to the level of technical detail. Some of the information on the barcoding
procedure and interpretation (e.g., the CBOL criteria, the description of the primers and sequence overlap,
the description of the result qualifier scores) are quite technical and largely inaccessible to even scientific
readers who are not previously acquainted with the terminology and procedures. Other parts (e.g.,
explanation of DNA structure and base-pairing) are much more straightforward and apparently aimed at a
different audience.
In the end, all of the method description and data presentation does not lead to a clear discussion of the
barcoding findings and their implications.
Specific Comments
Pg. 11, par. 2 - The text is unclear as to how duplicates and replicates were treated in terms of the value
for each composite sample that was actually entered into the overall statistical analysis.
Pg 12, par. 2 - The treatment of non-detects appears to have little overall impact on the estimation of the
average Hg concentration across species and on the use of these data for fish consumption advisories
(because the non-detects are in the species with the lowest Hg concentrations). It should, however, be
noted for the sake of completeness that the comparison between C = ND/2 and C = ND is a biased
comparison since it assumes that the true value of the concentration in the case of a non-detect can only
A-24
-------
be larger than the theoretical average (i.e., ND/2). A more balanced comparison would be among C =
ND/4, C = ND/2 and C = ND.
Pg. 21, par. 2 - The reference to large R2 values resulting from "large separations in data points" sound
like the effect of influential/outlying observations. If the concern is with such observations unduly
influencing the correlation, were non-parametric correlations considered?
Pg. 22 (equation) - The equation is not comprehensible as printed although it can be decoded from the
variable definitions below.
Pg. 31, par. 3 - "... and includes attributes of the level important to its meaning or implementation. " This
language is unclear.
Pg. 37, par. 3 - "Table D-la" The primary reference here should probably be to Table 14.
"Converting to meals per 30-day month as shown in Table 14" The reference here should be to
Table 14b.
"Using the conservative estimate ofHg concentration... also should not be eaten weekly. " The
issue at this point in the text is monthly consumption. Weekly consumption was already addressed.
Pg. 43, par. 2 - The format needs to be edited.
Par. 3 - The basis for CM-FDA matching should be clarified.
Table 15-1 don't understand why the species detail in the CM data present in Table 2 was not
carried forward here.
Pg. 45, par. 3 - "... the mean CM concentration... the only exception is bluefish" Swordfish also appears
to be an exception.
Pg. 117, par. 1 - "Two samples... selected only for mercury analysis were also analyzed for PCBs " I
don't understand this.
Pg. 119, par. 2 - "The waterbodies of origin for each sample are also shown, but no clear trends appear
to connect the PCB concentrations with the waterbodies of origin. " Another way to look at the data in
Table C-15 is that the farm-raised catfish were uniformly low in PCBs.
A-25
------- |