External Peer Review of EPA’s Draft Reassessment of the 

1999 Ambient Water Quality Criteria 

for Ammonia – Freshwater

Peer Reviewer Comments

Final Report

EPA Contract No. EP-C-07-059

Work Assignment No. 1-04

Submitted to:

Lisa Huff

Health and Ecological Criteria Division

Office of Science and Technology

U.S. Environmental Protection Agency

1200 Pennsylvania Avenue, NW

Washington, DC  20460

Submitted by:

Eastern Research Group, Inc.

110 Hartwell Avenue

Lexington, MA  02421-3136

September 21, 2009

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.

Table of Contents

SECTION A: Technical Charge to Peer Reviewers	A-1

SECTION B: Peer Review Comments Organized by Charge Question	B-1

  TOC \o "1-3" \h \z \u    HYPERLINK \l "_Toc239592396"  Acute criteria
in fresh waters	B-  PAGEREF _Toc239592396 \h  3  

  HYPERLINK \l "_Toc239592397"  1.	Are the toxicity tests used to derive
the criteria scientifically defensible for such use?  Are you aware of
other relevant data that were not used?	B-  PAGEREF _Toc239592397 \h  3 


2  HYPERLINK \l "_Toc239592398"  .	What are the technical considerations
that EPA should evaluate when mussels are present and mussels are absent
with respect to the recommended acute criteria?	B-  PAGEREF
_Toc239592398 \h  7  

  HYPERLINK \l "_Toc239592399"  3.	Is it scientifically defensible to
exclude the glochidia data at this time due to the uncertainty of
appropriate test duration time for this life stage?  Do you believe
there is an alternative approach to the use of this data that would be
more scientifically sound?	B-  PAGEREF _Toc239592399 \h  11  

  HYPERLINK \l "_Toc239592400"  4.	Regarding the proposed approach to
glochidia data in the 2009 draft position statement as it relates to
ecological relevance and practicality - Is the approach a scientifically
defensible principle for structuring the population exposure duration
problem and designing further research to quantify such a duration?	B- 
PAGEREF _Toc239592400 \h  15  

  HYPERLINK \l "_Toc239592401"  Hyalella azteca position statement and
proposed rationale	B-  PAGEREF _Toc239592401 \h  19  

  HYPERLINK \l "_Toc239592402"  5.	Are the position statement and
supporting rationale regarding use of toxicity data for Hyalella azteca
in criteria development reasonable and scientifically sound
recommendations?	B-  PAGEREF _Toc239592402 \h  19  

  HYPERLINK \l "_Toc239592403"  Chronic Freshwater Criterion	B-  PAGEREF
_Toc239592403 \h  21  

  HYPERLINK \l "_Toc239592404"  6.	Are the toxicity tests and other
studies used to derive the criterion scientifically defensible for such
use?  Are you aware of other relevant data that were not used?	B- 
PAGEREF _Toc239592404 \h  21  

  HYPERLINK \l "_Toc239592405"  7.	Is the freshwater chronic criterion
scientifically defensible with mussels present and mussels absent?	B- 
PAGEREF _Toc239592405 \h  25  

  HYPERLINK \l "_Toc239592406"  Use of 28-day Juvenile Test Data	B- 
PAGEREF _Toc239592406 \h  29  

  HYPERLINK \l "_Toc239592407"  8.	Given that the juvenile life stage of
freshwater mussels is relatively long (2-6 years) are 28-day exposure
tests with juvenile mussels scientifically defensible as "chronic" test
data for criteria development?	B-  PAGEREF _Toc239592407 \h  29  

  HYPERLINK \l "_Toc239592408"  9.	Should toxicity studies on the growth
rates of mussel shells during 28-day tests be considered quantitatively
when developing water quality criteria?	B-  PAGEREF _Toc239592408 \h  33
 

  HYPERLINK \l "_Toc239592409"  10.	 Regarding the position statement
and rationale on use of juvenile mussel growth data –	B-  PAGEREF
_Toc239592409 \h  35  

  HYPERLINK \l "_Toc239592410"  11.	The values of the acute and chronic
ammonia criteria have a strong dependence on pH.  Juvenile and adult
mussels, as sediment-dwelling organisms, inhabit a medium that may have
vertical pH gradients, thereby creating some uncertainty about the
appropriate pH to assign as their exposure conditions.  For applying a
criterion protecting mussels, do you have suggestions on how states and
EPA might determine the pH applicable to the sediment micro-environment
to which mussels are typically exposed?	B-  PAGEREF _Toc239592410 \h  39
 

  HYPERLINK \l "_Toc239592411"  12.	In general, should the criteria
include a consideration for the potential pH difference between sediment
and the water?  If so, what is the most scientifically defensible way to
account for these differences when deriving protective water quality
criteria?	B-  PAGEREF _Toc239592411 \h  43  

  HYPERLINK \l "_Toc239592412"  13.	Should exposure tests on juvenile
mussels be conducted with or without sediment in the test chamber?	B- 
PAGEREF _Toc239592412 \h  45  

  HYPERLINK \l "_Toc239592413"  Additional Comments	B-  PAGEREF
_Toc239592413 \h  49  

SECTION C:    HYPERLINK \l "_Toc239592415"  Individual Peer Reviewer
Comments	C-  PAGEREF _Toc239592415 \h  1  

  HYPERLINK \l "_Toc239592416"  Alex Barron	C-  PAGEREF _Toc239592416 \h
 3  

  HYPERLINK \l "_Toc239592417"  Steven P. Canton. MS	C-  PAGEREF
_Toc239592417 \h  17  

  HYPERLINK \l "_Toc239592419"  Jerome M. Diamond, Ph.D.	C-  PAGEREF
_Toc239592419 \h  31  

  HYPERLINK \l "_Toc239592421"  George Dixon, Ph.D.	C-  PAGEREF
_Toc239592421 \h  47  

  HYPERLINK \l "_Toc239592424"  Jerry L. Farris, Ph.D.	C-  PAGEREF
_Toc239592424 \h  53  

  HYPERLINK \l "_Toc239592426"  William L. Goodfellow Jr., MS	C- 
PAGEREF _Toc239592426 \h  63  

  HYPERLINK \l "_Toc239592427"  James J. Pletl, Ph.D.	C-  PAGEREF
_Toc239592427 \h  71  

 



  SEQ CHAPTER \h \r 1 Charge to Peer Reviewers

EPA’s Draft Reassessment of the 1999 Ambient Water Quality 

Criteria for Ammonia – Freshwater

July 24, 2009

The document to be reviewed is a draft revision of EPA’s aquatic life
criterion for ammonia.  The new criterion is intended to replace the
current aquatic life criterion previously updated in 1999.  

BACKGROUND:   

  SEQ CHAPTER \h \r 1 The aquatic life criterion for ammonia is used for
determining the level of pollution control necessary to attain the
aquatic life uses of water bodies, and in particular to allow the
propagation of fish and shellfish, per Section 302 of the Clean Water
Act.

In a 2004 interagency memo EPA stated that it would conduct a criteria
re-evaluation of the 1999 ammonia aquatic life criteria in light of new
studies indicating that freshwater mussels may be more sensitive to
ammonia than the organisms considered in the 1999 Update of Ambient
Water Quality Criteria for Ammonia.  In addition, a Federal Register
Notice (FRN) was published in 2004 notifying the public of the criteria
re-evaluation and requesting any pertinent toxicity data.  Responding to
comments from the FRN regarding concerns about uncertainty in the new
test protocol and data, in 2005 EPA hosted the Mussel Toxicity Testing
Protocol workshop to convene experts to discuss the new test protocol
for glochidia and juvenile mussels.  

The current draft criteria reassessment of the 1999 Update incorporates
new toxicity data on freshwater mussels as well as snails and a couple
of ESA listed fish species.  The focus of the 1999 Update was on the
chronic criteria whereas this reassessment updates both the acute and
chronic criteria.  The current draft recommendation includes criteria
for waters with mussels present and another applied to waters with
mussels absent.

  SEQ CHAPTER \h \r 1 DOCUMENT TO BE REVIEWED

The document is titled “Reassessment of the 1999 Ambient Water Quality
Criteria for Ammonia - Freshwater”.  The document presents acute and
chronic ammonia criteria for fresh waters with mussels present and a
separate criterion for waters with mussels absent. 



CHARGE QUESTIONS

Acute criteria in fresh waters:

1.	Are the toxicity tests used to derive the criteria scientifically
defensible for such use?  Are you aware of other relevant data that were
not used?

What are the technical considerations that EPA should evaluate when
mussels are present and mussels are absent with respect to the
recommended acute criteria?

3.	Is it scientifically defensible to exclude the glochidia data at this
time due to the uncertainty of appropriate test duration time for this
life stage?  Do you believe there is an alternative approach to the use
of this data that would be more scientifically sound?

4.	Regarding the proposed approach to glochidia data in the 2009 draft
position statement as it relates to ecological relevance and
practicality - Is the approach a scientifically defensible principle for
structuring the population exposure duration problem and designing
further research to quantify such a duration?

Hyalella azteca position statement and proposed rationale (see Appendix
B):

The EPA workgroup developed a position statement and proposed supporting
rationale describing the concerns over using Hyalella azteca toxicity
test data in criteria development due to the uncertain health of the
test organisms in different test water composition.  The rationale
defines the specific concerns and uncertainties supporting the
recommended exclusion of the Hyalella data from use in criteria
derivation, at this time; the position statement is based on the
workgroup’s review of a number of toxicity tests on Hyalella,
referenced in the rationale.  

5.	Are the position statement and supporting rationale regarding use of
toxicity data for Hyalella azteca in criteria development reasonable and
scientifically sound recommendations?  

  SEQ CHAPTER \h \r 1 Chronic Freshwater Criterion:

6.	Are the toxicity tests and other studies used to derive the criterion
scientifically defensible for such use?  Are you aware of other relevant
data that were not used?

7.	Is the freshwater chronic criterion scientifically defensible with
mussels present and mussels absent?  



Use of 28-day Juvenile Test Data (see Appendix C):

Water quality criteria for the protection of aquatic life are derived
using toxicity endpoints that relate to population level impacts.  In
general, these endpoints relate to survival, growth and/or reproduction.
 The 28-day test with juvenile mussels, while similar in duration to a
standard chronic test, is not technically an early-life stage test
according to the 1985 Guidelines for Aquatic Life Criteria as much of
the early development will have already occurred.

8.	Given that the juvenile life stage of freshwater mussels is
relatively long (2-6 years) are 28-day exposure tests with juvenile
mussels scientifically defensible as "chronic" test data for criteria
development?   

9.	Should toxicity studies on the growth rates of mussel shells during
28-day tests be considered quantitatively when developing water quality
criteria?

10.	 Regarding the position statement and rationale on use of juvenile
mussel growth data – 

Is it scientifically defensible to include the juvenile growth data from
a 28-day exposure period as “other data” that might influence the
criteria however not be used directly in the derivation of the criteria
value? 

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

11.	The values of the acute and chronic ammonia criteria have a strong
dependence on pH.  Juvenile and adult mussels, as sediment-dwelling
organisms, inhabit a medium that may have vertical pH gradients, thereby
creating some uncertainty about the appropriate pH to assign as their
exposure conditions.  For applying a criterion protecting mussels, do
you have suggestions on how states and EPA might determine the pH
applicable to the sediment micro-environment to which mussels are
typically exposed?

12.	In general, should the criteria include a consideration for the
potential pH difference between sediment and the water?  If so, what is
the most scientifically defensible way to account for these differences
when deriving protective water quality criteria?

13.	Should exposure tests on juvenile mussels be conducted with or
without sediment in the test chamber?  

 

SECTION B

Peer Review Comments Organized by Charge Question

Acute criteria in fresh waters

1.	Are the toxicity tests used to derive the criteria scientifically
defensible for such use?  Are you aware of other relevant data that were
not used?

Alex Barron	Yes, all the toxicity tests that influence the derivation of
the new criteria for ammonia appear to be scientifically defensible. All
relevant data that I am aware of has been considered in this
reassessment document.

Steven Canton	Yes, in general the toxicity tests used to derive the
water column criteria, as cited in the draft update, are scientifically
defensible.  Plus, it appears they provide sufficient data to evaluate
inter and intra-species comparisons using the guidelines established for
deriving water quality criteria.  

Additionally, a few years ago, we conducted a review of a number of
ambient water quality criteria (AWQC) as part of the Arid West Water
Quality Research Project (AWWQRP 2006).  One aspect of that project
included a review and update to the 1999 EPA Ammonia criteria document,
similar to this draft.  Checking those updates side-by-side, it appears
there are a few more studies that could be included in this
re-evaluation:

Augspurger, T., A.E. Keller, M.C. Black, W.C. Cope, and F.J. Dwyer.
2003. Water quality guidance for protection of freshwater mussels
(Unionidae) from ammonia exposure. Environmental Toxicology and
Chemistry 22:2569-2575.

Khatami, S. H., D. Pascoe, M. A. Learner.  1998.  The acute toxicity of
phenol and unionized ammonia, separately and together, to the
ephemeropteran Baetis rhodani (Pictet). Environmental Pollution
99:379-387.

Rubin, A.J., and M.A. Elmaraghy. 1977. Studies on the toxicity of
ammonia, nitrate and their mixtures to guppy fry. Water Research
11:927-935.

Russo, R.C., D.J. Randall, and R.V. Thurston. 1988. Ammonia toxicity and
metabolism in fishes. In R.C. Ryans, ed. Protection of River Basins,
Lakes, and Estuaries. pp. 159-173. American Fisheries Society, Bethesda,
Maryland.

Sangli, A. B. and V. V. Kanabur. 2001. Toxicity of ammonia to a
freshwater fish, Gambusia affinis and its effect on oxygen consumption.
Geobios 28:56-58.

Tomasso, J.R., C.A. Goudie, B.A. Simco, and K.B. Davis. 1980. Effects of
environmental pH and calcium on ammonia toxicity in channel catfish.
Transactions of the American Fisheries Society 109:229-234.

I should also note that in that prior review of the 1999 Ammonia
document (AWWQRP 2006), we raised concerns regarding the appropriateness
of some studies/data used in the 1999 EPA document (see Table 1 below). 
These studies may need an additional evaluation to determine if they
provide toxicity data meeting EPA data quality guidelines (Stephan et
al. 1985). 

Table 1

Deletions of Inappropriate Data Presented in the EPA 1999 AWQC Ammonia
Document (from AWWQRP 2006)

Species

References

Comments

Acute

Simocephalus vetulus

Mount 1982

Insufficient data to validate results

Oncorhynchus mykiss

Calamari et al. 1977

Unable to validate citation or data-other data available by same author
with this species

Oncorhynchus mykiss

Thurston et al. 1981

Unable to validate citation or data

Oncorhynchus mykiss

Reinbold & Pescitelli 1982

Unable to validate citation or data

Pimephales promelas

Thurston et al. 1981

Unable to validate citation or data

Pimephales promelas

Reinbold & Pescitelli 1982

Unable to validate citation or data

Gambusia affinis

Wallen et al. 1957

Insufficient data to validate results

Lepomis macrochirus

Reinbold & Pescitelli 1982

Unable to validate citation or data

Sander vitreum

Reinbold & Pescitelli 1982

Unable to validate citation or data

Chronic

Ictalurus punctatus

Colt & Tchobanoglous 1978

Insufficient data to validate results

Source: (AWWQRP 2006)



Jerome Diamond	In general, the toxicity tests used to derive ammonia
criteria in the draft Reassessment Document (RD) are scientifically
defensible given EPA’s criteria development guidelines (Stephan et al.
1985).  However, I have reservations regarding the use of snail data
derived from field-collected organisms as explained below under Chronic
Freshwater Criterion.  I am not aware of other relevant freshwater
toxicity test data that were not either used or considered in the RD.

George Dixon	I am familiar with the majority of the literature cited
and, in my opinion, the tests used to develop the acute criteria are
scientifically defensible. I am, however, relying on the fact that the
Draft Reassessment states that a process to identify data meeting the
minimum requirements for inclusion was followed.  I have not obtained
all of the cited literature and reviewed each paper for suitability of
inclusion of the findings in the data sets; this type of detailed audit
is beyond the scope of this review. I follow the ammonia literature
closely, and I am unaware of any additional data that could be of use.

Jerry Farris	With greater recent attention dedicated to standardization
of test techniques suitably addressing the feasibility of repeatability
and precision of results from juvenile mussel responses (Augspurger,
2007), data available at the time of this reassessment meet a higher
standard of acceptability for guidance modified from the US EPA (Stephan
et al., 1985) and follow consensus test protocols and quality assurance
recommendations sufficient to support water quality criteria
development.   The majority of tests used to derive the criteria seem
scientifically defensible for such use.  Upon review of those included
and others not specifically included in the calculation, but cited in
Table 1 of the draft addendum, there are tests that present concern for
acclimation procedures of test organisms and culturing (either reported
or unreported) and lack of reference testing that can insure level of
expertise or consistency associated with such field collection, organism
transfer, and culturing that can significantly influence test responses.
 Specifically, sufficient acclimation information for water quality and
temperature was not included in Mummert et al., 2003 and Scheller, 1997.
 

I am not aware of other data from standardized toxicity tests with
ammonia that could be considered relevant to this reassessment.  

Bill Goodfellow	From my review, I believe that the acute toxicity
testing data is relevant. My only concern is that this document was
careful to use data that was from published references which I believe
is a strength of this document as compared to earlier water quality
criteria documents. Since the only unpublished data used that
demonstrates a somewhat low toxicity response to ammonia was from Keller
1999 for the Pondshell Mussel, maybe putting a synopsis as an attachment
would be warranted.  This would help make the document transparent as
the other Attachments help do.  

The Table A in text of the GMAV for the Mountain Whitefish does not
match Table 1. The Text table reports 12.11 and Table 1 reports 12.09. 
I am expecting that the Table 1 is correct and is what was used for the
FAV. Please verify.

Jim Pletl	The question requires that the term “scientifically
defensible” be defined within the context of the use or the
consequences of the use.  The use is to support development of water
quality criteria that will be promulgated by states as legally
enforceable water quality standards. Failure to meet water quality
standards instream and/or at the edge of respective mixing zones will
result in TMDLs, implementation plans, or NPDES permit limits for
dischargers to the respective stream.  Failure to meet water quality
standards also infers that designated uses are not being supported.  In
the case of ammonia criteria, the designated use is aquatic life.  The
Clean Water Act also states that violations of such permit limits can
result in significant monetary fines and/or jail time for responsible
individuals.  Therefore the use, in this case, is quite demanding of the
data.  

The level of scientific defensibility set by EPA for this use is limited
to that defined by EPA’s 1985 “Guidelines For Deriving Numerical
National Water Quality Criteria For The Protection of Aquatic Organisms
And Their Uses”, hereafter referred to as the Guidelines, and the
published, peer reviewed toxicity test methods used in each case.  The
tests in question appear to meet these standards; however the data
quality objectives set by these documents may not be sufficient for the
data use.  The toxicity test methods and the Guidelines do not require
what are considered to be critical elements of scientifically defensible
toxicity tests.  Acute tests typically only have one quality objective:
minimum survival in controls.  Other than control survival there are no
other quality objectives that must be met for the results of these tests
to be considered reliable.  For example, 1) the methods and Guidelines
do not require a concentration-response curve for data to be deemed
acceptable for any use.  There can be little certainty in a toxicity
test result where response does not change predictably with the
concentration exposure.  If demonstration of this curve is not required
of the test or EPA, data of questionable quality may be used to derive
water quality criteria.  2) Analysis of acute toxicity test data
typically also does not address the intra-treatment variability that
commonly occurs in these tests; providing results with biased estimates
on uncertainty.  3) The Guidelines also do not require replication of
any test result for the most sensitive species.  This would allow a
single test result to significantly bias the calculation of water
quality criteria.  ASTM E2455-06 for testing freshwater mussels
recommends that “different batches of the same species and the same
life stage should be collected and tested over time in order to obtain a
measure of the variability associated with testing the particular
species”.  This goal was not accomplished for the pink mucket, the
third most acutely sensitive genus of the database.  

Aside from universal issues for acute tests 4) there is also concern
that the acute juvenile mussel tests may be biased towards findings of
more sensitivity and toxicity.  The tests proposed to be used for this
reassessment are conducted without sediment in the test vessels.  Newton
and Bartsch (2007) found that without sediment juvenile organisms did
not grow in 96 hour toxicity tests.  This clearly shows that the
organisms are stressed by the lack of sediment in the test vessels. 
This also strongly suggests that the survival endpoints from acute tests
may be biased low (falsely indicating greater toxicity) due to the added
stress of failing to provide sediment in the test vessels.  All of these
issues characterize a significant level of uncertainty which is
indirectly proportional to scientific defensibility.  EPA should address
these uncertainties before using new data to update this criteria
document.

I am not aware of other relevant data that were not used for this
reassessment.

What are the technical considerations that EPA should evaluate when
mussels are present and mussels are absent with respect to the
recommended acute criteria?

Alex Barron	Table B compares Final Acute Value (FAV) and Criterion
Maximum Concentration (CMC) values calculated with the entire dataset
including data for bivalve mollusks and also with the dataset with the
data for bivalve mollusks removed.  The Asiatic clam (Corbicula
fluminea) is in a different Family (Cobiculidae) and Order (Veneroida),
and is not a Unionid mussel.  Corbicula has shown very similar
sensitivity to ammonia compared to Unionid mussels however.  Table B
really represents a non-bivalve mollusk dataset, which would be
appropriate where no bivalve mollusks of any kind are present.  The FAV
and CMC calculated without the data for bivalve mollusks is less than a
factor of two different than the FAV using the entire dataset, which is
62.75 % higher.   The 1999 ammonia criteria also provided separate
recommended acute criteria based on observed sensitivity of salmonid
species.  The 1999 criteria document also presented separate acute
criteria that are approximately 50 % different between the salmonids
present or absent acute criteria. The recommendations in the 2009
reassessment document providing separate criteria , calculated with and
without a group of sensitive species is in keeping with precedents set
in the earlier ammonia criteria for calculating separate criteria when
one group of species appears to be especially sensitive to ammonia.   
This can be considered appropriate for ammonia criteria because of the
relatively large database and the fact that ammonia toxicity is usually
caused by mg/L concentrations as compared to ug/L concentrations as seen
with many other toxic pollutants. 

The implementation of criteria based on the dataset without freshwater
mussels will require some way of determining whether or not a water body
is not likely to contain freshwater mussels, before such criteria can be
considered appropriate.  The presence or absence of freshwater mussels
in a water body may be difficult to ascertain without a survey that
specifically targets this issue.  Most monitoring techniques for
collecting benthic macroinvertebrates such as the widely used EPA Rapid
Bioassessment Protocol (RBP) are not designed or intended to regularly
collect mussels that are often tightly imbedded in the sediment.  Thus a
typical RBP survey that did not collect freshwater mussels cannot be
accepted as proof of absence of mussels in that waterbody.  I would
recommend a default of assuming the presence of freshwater mussels
unless several surveys that specifically targeted the identification of
mussels have been conducted on that water body.  A brief review of
biological survey data in Virginia showed that the presence of
freshwater mussels was detected in 83 % of water bodies that had been
surveyed with methods capable of finding mussels.  However, in some
waterbodies the presence of mussels was confirmed only after several
surveys. Because the distribution of freshwater mussels is often
dependent on fish for dispersal, it might be suspected that mussels may
not be present in some headwater or intermittent steams where it can be
demonstrated that fish are not present due to physical barriers
preventing fish migration.  

Steven Canton	Interestingly, we made a similar recommendation for this
particular situation (i.e., the concept of with and without
mussels-based criteria for ammonia) in that prior report (AWWQRP 2006,
Chapter 4).   I have attached a copy of the relevant chapter from that
analysis, which may be useful, as least in reference to how this has
been looked at by other parties. 

In addition to a with and without mussel approach, the document could
also include other “implementation” issues, such as evaluating
additional site-specific information needed to provide in-depth
technical recommendations.  For example:

Is there a definitive cause and effect relationship between ammonia
sources and the diversity and abundance of mussels downstream of the
source?

How have the manageable sources (i.e., Agriculture, WWTP) of nitrogen
changed over the last decade within the study waters in comparison to
measured changes in mussel communities?  

Is there evidence of anthropogenic nitrogen loading that is equal to or
greater than the natural occurrence of decomposition due to
heterotrophic bacteria?

What is the within-site spatial and temporal variability in NH3 of the
pore water and is this variability linked to season, storm events,
temperature, or flow conditions.

In addition it would be advisable to include a statement to recognize
the importance of developing “site-specific standards”, since
mussels now play an important role and are the most sensitive species in
the list. Site-specific standards reflect the reality of which species
are needed to be protected in a particular waterbody, particularly with
mussels that may be more abundant in some parts of the country, whereas
in some other regions their presence can be very limited.  In this case,
some guidance on when and how to do mussel surveys to determine when to
apply a “with mussels” or “without mussels” criterion would be a
useful addition to the criteria document.

Jerome Diamond	Attempting to differentiate freshwater ammonia criteria
on the basis of mussels being present or absent is in my opinion a very
tenuous and probably unsupportable idea scientifically.  First, the
acute data indicate that Corbicula (Asiatic Clam) is nearly as acutely
sensitive as the most sensitive unionids.  Corbicula is not a unionid
and has a totally different life cycle than unionids.  Furthermore, they
are now widespread in the U.S. so it may prove fairly inefficient to
separate criteria on the basis of presence of bivalves in general.  

As an implementation issue, distinguishing between surface waters with
mussels and those without is not nearly as straightforward as it is for
salmonids or “cold water” species presence/absence because the
latter are dependent on a fairly known, measurable water characteristic
(water temperature regime).  Also, while most states have cold water as
well as warm water aquatic life uses in their water quality standards,
I’m not aware of many (perhaps any) that have specifically a mussels
use.  The only example of which I’m aware is Virginia DEQ, which
distinguishes different application of the chlorine standard based on
whether federally listed mussels are known to be present in a given
stream segment. They apply a “halogen ban” (no chlorine or other
halogenated disinfection chemicals are allowed to be used for wastewater
treatment) in those segments for those permittees that do not have
intermittent dischargers and have discharges <  20,000 gallons per day. 
 However, in this case the water quality standard was not changed but
rather a risk management approach was taken to reduce the major point
sources of chlorine or other halogenated disinfection chemicals to the
stream.

Mussels can and do occur (or have occurred) in a wide variety of habitat
types and water quality regimes and so their presence is not often known
without specifically sampling for them.  Such sampling may need to be
fairly intensive to “prove” the absence of mussels.  In addition,
mussels have been extirpated from a number of streams in the U.S. due to
historic land use and habitat changes, toxic spills, and poorly treated
wastewater and nonpoint runoff (e.g., see EPA 2002).  If today, a state
finds that mussels are not present but historically they were, what then
should be the ammonia criteria?  I believe such differentiations in
criteria suggested in the RD be relegated to EPA’s recalculation of
site-specific criteria (USEPA 1994).  That procedure was designed to
identify whether certain types of species are present or could occur in
a given site or waterbody.  This would be a more efficient and
scientifically defensible way to handle the apparent sensitivity of
freshwater mussels to ammonia.

George Dixon	The document is silent on the appropriate methods for
determining whether mussels are present or absent in a given receiving
environment, and therefore which criteria should be applied. I expect
that a guidance document will have to be developed to assist those
applying the revised criteria. My expertise is in aquatic toxicology,
while the question of presence or absence is based more in field-based
aquatic ecology, particularly mussel ecology.    As such I do not have
detailed recommendations for the determination. I would, however, expect
that some consideration would be given to a method that was able to deal
with environments where mussels were once present, but have now been
expatriated. 

Jerry Farris	With mussels presenting such an extreme sensitivity and
complex life history in comparison to other groups, not only will their
presence establish consideration for site variance demonstrations
related to specific media, habitat or host fish interactions, but their
spatial positions and ecological requirements against the backdrop of a
stream’s ability to meet water quality standards- will require some
unique approaches to:

Short- term authorizations to discharge

Compliance with permits that account for ammonia levels, but in streams
or those segments that fail to meet water quality standards

Designated mixing zones and how “shock effect” associated with
ammonia will be considered 

The applicability of biocriteria and use of in-stream community
assessments related to the concept of mussels present and absent that
would incorporate differences between degree of colonization and
distribution rates for fish, other benthic organisms of conventional use
in stream surveys, and freshwater mussels.   These considerations
coupled with only a recent attempt to describe how certain stream
characteristics may serve to structure mussel communities, will need to
be carefully approached when and if such recommendation is made.  

 

Since many of the ‘mussel present’ instances may involve listed
species, the ability to qualify broad testing and culturing of suitable
surrogate species for the time and seasonal limits related to any
testing related to permit or enforcement actions, will likely present an
additional set of laboratory qualifying or certification considerations
related to accessing suitably sited mussel species for testing. 
Laboratories listed in reference to the range of studies cited with this
reassessment and that have implemented the new test standards (ASTM,
2006) have certainly advanced the methodologies needed for risk
assessments and management decisions.  However, this relatively limited
number of research and agency laboratories will likely be challenged to
fulfill the rigorous requirements of consistency with organism culturing
and quality assurance that may be expected of more conventional test
demands.  While consideration of these outcomes may reach beyond the
scope of the question for mussel presence, increased demand as a
technical consideration would be expected to impact population
availability.  

Relative use of reference sites in determining relationships between
areas of reduced freshwater mussel diversity and abundance and known
significant ammonia sources will also likely become a technical
consideration.  Definitive cause-and-effect relationships have not been
documented in such instances (Bartsch et al., 2003) and such
considerations would require more detailed attention between the field
exposure and effects data.   

Bill Goodfellow	I believe that EPA has adequately defended their
position for when mussels were used or not used for the generation of
the two phased acute criterion.  

I do feel that EPA should include guidance that the lower mussel
criterion should only be used when mussels are currently present in the
water system and not historic or wished to be in the system. I can see
States using the lower mussel numbers without a strong technical reason
to use them because someone wished them to be in the system, etc.  When
States choose to go this way arbitrarily, it will weaken the
acceptability of the criterion for everyone.

Jim Pletl	Generally, water quality criteria are designed to protect
populations.  The first step in determining whether mussels are present
or absent is to define a standard for the conditions describing that a
population exists at a site.  This is not a straight-forward exercise
and can be quite difficult as seen in various site-specific criteria
projects around the country.  There must be guidance as to what
parameters and measures of those parameters are to be used to define
presence/absence of a population.  Quantitative evidence of reproduction
would be required at a minimum, as well as the presence of various
developmental stages.  The normal abiotic conditions documented for a
species should also be compared to that of a particular surface water to
determine whether a conclusion of presence or absence is supported. 
Certainly the availability of host species could influence the
determination of absence/presence.  

Once absence/presence is established EPA should only use mussel data to
develop water quality criteria for waters where mussels are deemed
present.  EPA should ensure that conditions of the site where absence
has been concluded have not been anthropogenically changed to a state
where mussels cannot reproduce.  Historic records of water quality and
biological surveys would help support this determination.  It is clear
that EPA and the states will share the burden of the presence/absence
determination and that there will be differences of opinion as to
whether, when absence seems apparent, conditions changed due to human
activity that could have been reasonably precluded or the species simply
never successfully established a population at the site due to
species-specific factors.  

EPA should also determine whether simply the presence of mussels is
important or the presence of specific species is important.  Given that
the sensitivity of mussel species in acute tests spans a six-fold
concentration range the presence of only certain species could have a
similar effect on the resulting water quality criteria; which then could
have significant impacts on water quality management decisions. 
Further, different species have different life cycle durations.  If
individual species were used to develop criteria rather than the
presence/absence of a taxonomic group the differences in life cycles may
also impact criteria implementation.  

3.	Is it scientifically defensible to exclude the glochidia data at
this time due to the uncertainty of appropriate test duration time for
this life stage?  Do you believe there is an alternative approach to the
use of this data that would be more scientifically sound?

Alex Barron	The position statement in Appendix B of the reassessment
document acknowledges several concerns about the use of toxicity data
for glochidia and recommends that these data should not be used for
criteria development at the present time.   This position seems
reasonable given the uncertainties regarding glochidia testing as
outlined in Appendix B. Glochidia, with such short-lived life stages and
with variable lengths of natural viability in the water column, ranging
from a few hours to several days, it is difficult to know how to use
some of these data for developing criteria.   For some species where the
natural viability of glochidia, or where the actual exposure to the
water column is short, the result of 48 hour tests may not provide a
realistic exposure scenario, while a shorter exposure may be more
important..   

There are some additional short-term (6 and 24-hour) exposure data
available that may provide some additional information on this issue.
Table 2 of the Reassessment Document provides acute values based on
48-hour exposures for tests conducted with glochidia in several
publications.   For some of these same tests, additional information on
6 hour and 24-hour exposures are also contained in Wang et al 2007b, in
their table 3.  These 6 and 24-hour data indicate that glochidia are
less sensitive to ammonia during 6 and 24-hour exposures than during a
48 hour exposure. For the species where glochidia may only be viable or
exposed to the water column for a few hours, the results of 6 hour and
24 hours tests may provide useful information and these data could be
normalized to the proper standard conditions and included in Table 2, to
provide additional information for comparison.  As reported in Wang et
al 2007b; the data suggest that at 6 to 24 hour test durations, the
sensitivity of glochidia for most species of freshwater mussels is less
than or equal to the sensitivity of the juveniles  of the same species
at 96 hours.   This provides some support that the revised ammonia acute
criteria (based on 96 hour data for juveniles) may also provide
protection to glochidia at exposure durations of 24 hours or less.   The
fact the acute criterion is applied as a one-hour average should also
provide some additional level of protection.

Steven Canton	Given the rationale provided, in combination with EPA
criteria development guidance, general ecotoxicological principles, and
my own views on general invertebrate biology, I agree it is
scientifically defensible to exclude glochidia data in ammonia criteria
development for the following reasons. 

First, the document notes that despite having  four acute toxicity
studies that consider five different mussel species (with the glochidia
stage) ,with the majority of test having a duration of 2 days, there are
still several questions to be answered when testing these particular
life stages:

How long do they stay attached to the host?

How long do they stay in conglomerates and does this jelly (for those
species) protect them from exposure to contaminants?

If they sink to sediment how long do they stay there in the conglomerate
before transforming to the juvenile mussel?  

I agree with the authors that these valid questions and also note that
the questions will probably have different answers for different species
of mussels, which makes the data of even more uncertain general
ecological relevant and, therefore, use of glochidia data even more
problematic and difficult to apply generally to this group for criteria
development purposes.

Second, concerning the mechanisms of ammonia toxicity in this particular
stage of development there is also a long list of unanswered questions:
would glochidia be able to take up ammonia if available in the host?,
what is the mechanism of glodichia exposure to ammonia when they are
released in conglomerates?  - Again – great uncertainty with unknown
ecotoxicological significance.

Third, based on the results from those four studies it seems that if the
glochidia acute data were included in the calculations, they would
dramatically reduce the acute criteria and probably be over conservative
for other fresh water species without actually knowing if the
information has any ecological relevance.   

Given these uncertainties – which are well articulated in the draft
document, there appear to be two scientifically defensible approaches:
1) Do not use the glochidia acute toxicity data 2) Do a final acute
criteria calculation “with and without glochidia”. The first
approach would be the easiest to avoid confusion and possible errors in
development of safe levels for ammonia.  The second approach would give
flexibility – although it would be necessary to include a strong
caveat to a “with glochidia” value, given the questions noted in the
criteria document and my thoughts above. 

Jerome Diamond	I agree with the RD that glochidia data should not be
included in criteria derivation at this time but not entirely for the
reasons given by EPA.  First, a female mussel often produces hundreds or
even thousands of glochidia in a single season.  This helps counteract
the fact that a high percentage of the glochidia will not survive
because they have not encysted on a host fish within a reasonable amount
of time.  As many researchers have documented, glochidia are incapable
of surviving and maturing on their own—they require nourishment from a
host organism typically within days of being produced.  It could be
argues that basing criteria on a 50% reduction in glochidia survival
would be similar to basing criteria on a 50% reduction in egg survival
of other species that broadcast an abundance of eggs (e.g., marine
barnacles) in which only a small fraction of the larvae are likely to
survive and become established under natural, unstressed conditions.  I
am not aware that EPA has ever based acute criteria on egg survival
(even fertilized egg survival) and I believe this might contradict
EPA’s criteria development guidelines for establishing acute criteria.

A second reason I don’t think glochidia should be included at this
time is that the majority of mussel species at risk today (i.e.,
threatened, endangered, or species of concern) appear to have relatively
narrow specificity in terms of a vertebrate host and have evolved
structures (e.g., lures) or behaviors that enable very rapid encysting
in the host (e.g., see Rogers-Lowery and Dimock 2006).  Of the mussel
species actually used in glochidia testing that are discussed in the RD,
most of them have evolved adaptations to enable rapid infestation of a
host (often < 6 hours), resulting in a very short free-living glochidia
stage.  Thus, for the majority of mussel species at risk, exposure
periods > 24h are probably inappropriate, regardless of whether control
survival is > 90%.

A third reason is that while the ASTM glochidia test procedure appears
to be reasonably robust in terms of intra and inter-laboratory testing,
the procedure is still a “Guide”, not a “Method”, which
indicates that there are still many aspects requiring expert judgment,
trial and error, and research, especially in terms of culturing and
maintaining glochidia  under laboratory conditions.  Also the inter-lab
study reported by Wang et al. (2007) used the same water source and
glochidia stocks for all labs.  This would tend to underestimate the
true interlab variability, where each lab uses their own source of
glochidia (and techniques for obtaining glochidia) as well as water
source.  In addition, at the present time there are still relatively few
laboratories that have performed this test (e.g., compare with daphnia
or amphipod tests).  Much of the glochidia data under consideration were
produced by a few laboratories.  This may be acceptable for tests that
use species closely resembling those for which a standard EPA or ASTM
method has been developed and used extensively; e.g., an acute test with
a different species of minnow or Daphnia but the identical test design
and organism life stage as used for the standard fathead minnow or
Daphnia acute test.  However, such is not the case with the ASTM
glochidia test, which uses a fairly different test design and the
endpoint is dependent on sensitivity to a sudden stress (i.e., rapid
valve closure to NaCl exposure).  While the ASTM procedure appears to
address these concerns for the most part, there is still some
uncertainty in my mind regarding defensibility of the test endpoint.  As
noted in the ASTM method, exposures beyond 24h may not achieve 90%
control survival, indicating the precarious nature of keeping this
lifestage alive without a host.  Given this, is the apparent increase in
ammonia sensitivity between 6 and 48h a true difference in sensitivity
or is it an artifact of the method (i.e., glochidia are under increased
stress).  I also question the ecological importance of having valve
closure within one minute or less to the introduction of a NaCl
solution.  I’m not altogether sure this endpoint is similar to other
acute endpoints relied upon by EPA in their water quality criteria.  In
addition, the ASTM Guide points out that live or dead mussels could be
open or closed at a given time and some may respond more slowly to the
NaCl shock than others (Kernaghan et al. 2005).  If there was a more
direct way to determine mortality of glochidia to ammonia exposure, for
example, that would be preferable in terms of using such data for
criteria development.

I wonder if better use of glochidia survival data couldn’t be made by
considering the level of juvenile recruitment necessary to maintain a
viable mussel population, similar to the EPA fish larval recruitment
model for dissolved oxygen in coastal waters in the Virginian Province
(USEPA 2000).  Perhaps a similar type of framework could be used based
on the mussel species (or, if feasible and defensible, genus or even
family), when glochidia are typically produced in a given region, and a
function specifying ammonia toxicity to glochidia given ambient
temperatures and pH at that time.  The modeling framework could be made
general so that a user could input necessary temperature and pH data,
recruitment timing (e.g., season, month) and certain assumptions
regarding host abundance.  This would not be unlike the larval
recruitment model EPA has developed for dissolved oxygen with the
exception that the abundance of another species (i.e., a host) is not
needed or relevant for the dissolved oxygen model.  However, if such a
model was considered (albeit simplistic but using conservative
assumptions based on expert malacologist judgment), it could put
glochidia toxicity testing results into an ecological context in a more
useful way.

George Dixon	Exclusion of the glochidia data, as detailed in Appendix A,
is fully appropriate and well explained.  Until tests are completed to
determine (on a species by species basis) the duration of the period of
longevity (and by inference viability) for glochidia, LC50s obtained
with those life stages will be suspect. I am unaware of any other
approach which would be more scientifically valid 

Jerry Farris	Information from tests with glochidia can be supportive of
the consideration for site variance information, but should not be used
to provide stand alone estimates in the absence of supportive acute
juvenile estimates.   Multiple lines of evidence comprised of testing
with more than one mussel life stage that include comparisons with
glochidia responses, would seem more scientifically sound, and may
remove much of the seasonal dependence upon culturing and provision of
test organisms with less invasive techniques in instances mentioned in
the above mentioned technical considerations

Bill Goodfellow	Not only is it scientifically defensible, I believe that
it is the right way to go. Too much is unknown with regards to the
biology of glochidia and the associated toxicology or appropriate
methods to evaluate the toxicity of various chemicals for this life
stage.

Jim Pletl	The alternative question would be: Is it scientifically
defensible to include glochidia data at this time due to the uncertainty
of appropriate test duration time for this life stage?  Again, EPA has
not defined standards in the Guidelines for this determination other
than requiring acute tests be at least 48 hours in duration and test
organisms not be fed (except for mysids).  These requirements were not
developed to address the duration of the lifestage unique to glochidia. 
Toxicity data for glochidia appear to show that duration of tests with
this lifestage significantly affect the test results when tests are 24
hours in duration or longer.  Therefore it seems that the glochidia are
stressed in some fashion when tests meet the 48 hour requirement of the
Guidelines.  The test design for water quality criteria development is
intended to measure stress due to the treatment (ammonia) independent of
all other stressors.  Tests of 48 hour duration and more fail to meet
this goal.  Since tests of duration less than 48 hours do not meet the
Guidelines’ requirements but tests of duration equal to or greater
than 48 hours provide biased results it seems that a new standard for
test duration unique to these mussels must be first developed.  However
the literature indicates a great deal of diversity among freshwater
mussels in terms of their glochidia ecology.  The extent of this issue
demands either more detail in this criteria document or more
opportunities for site-specific ammonia criteria implemented by the
States for only the mussel species present.  The States will also
require guidance on how to develop site-specific ammonia criteria based
on the issues unique to glochidia.  To support the scientific
defensibility of EPA’s decision it would be appropriate to first
develop guidance for the appropriate test duration to be used for
developing water quality criteria relative to the duration of the life
stage.  Currently acute test duration commonly is within an order of
magnitude of life stage duration.  To accomplish this goal for glochidia
a test duration of 12 hours or less may be necessary.  Such a guideline
would help defend the scientific defensibility of using glochidia data
for developing water quality criteria.  

One should conclude that EPA’s decision to not use glochidia data is
logical and correct without more information. There is a great deal of
uncertainty associated with test results using glochidia.  The question
that must be answered is whether the level of uncertainty realized by
using glochidia data is so great that it compromises the integrity of
the resulting water quality criteria.  This, in fact, seems the case;
however it should be stressed that scientific defensibility should not
be based on judgment but minimum standards of quality.  Without these
standards of quality the magnitude of uncertainty is unknown and
probability of erroneous actions is elevated.

4.	Regarding the proposed approach to glochidia data in the 2009 draft
position statement as it relates to ecological relevance and
practicality - Is the approach a scientifically defensible principle for
structuring the population exposure duration problem and designing
further research to quantify such a duration?

Alex Barron	I believe that the approach outlined in appendix B is
defensible.  Given the significant uncertainties concerning the
appropriate exposure durations for glochidia as discussed in the
appendix B, attempting to use toxicity data currently available for
glochidia to derive defensible criteria at the present time appears to
be untenable.  This would be especially important if the final criteria
were significantly influenced by glochidia data with the recognized
uncertainties about practical ecological importance.  Water quality
criteria should be based on the best information available in order to
gain acceptance and avoid continued uncertainty.  Additional information
as discussed in appendix B would be needed to provide sufficient
confidence in utilizing glochidia data to derive water quality criteria.
 

Steven Canton	The proposed approach will give researchers and scientific
community time to come up with some answers and provide relevant
information concerning the appropriate time of duration for glochidia
exposure testing – if such testing is determined to be necessary for
this short-lived, parasitic life stage.  However, if this approach is
proposed, it will also create a limbo period, in which this issue will
have to be addressed while decisions are made for application of a
criteria that some may fight as “non-protective” since the criteria
document has “acknowledged” that a sensitive life stage was not
considered!

I’ll be the first to admit it is a difficult to task to structure a
criterion that includes population exposure approach to criteria
development when dealing with a wide variety of fresh water mussel
species that have different life spans (few months to several years),
not to mention a glochidia life-stage that can be released individually
or in conglomerates, and that can get attached to fish, rocks, sediment
or plants. 

Such variability could mean that for the particular issue of including
glochidia data or not, we perhaps should be treating these mussels
individually, by genus or species, rather than thinking of them simply
as the “mussel” group.  Regardless, it will take a considerable
amount of time and effort to develop some way to make “uniform”
relevant exposure toxicity testing time for the glochidia life stage
that will be relevant for all if not the majority of fresh water
mussels.

Jerome Diamond	I am not sure I concur entirely with the proposed
approach for further research discussed in Appendix A of the RD.  I do
agree that it would be useful to have better information regarding the
natural life expectancy of free-living glochidia in nature for several
representative species covering the range of different reproduction
strategies observed.  Such information would help indicate whether a 6,
24, or 48h exposure is warranted.  I do not, however, understand EPA’s
proposal specifying the free-living duration based on 95% of the
glochidia that attach to a host.  This appears to be a fairly
restrictive proposition and it is not clear to me how this would be
determined.  As Rogers-Lowery and Dimock (2006) observed, the
encapsulation process in fish varies with the species and organism
history of exposure to mussel infestations.  Many researchers have
demonstrated poor encapsulation rates of glochidia of certain mussel
species (including several listed as threatened or endangered) with
several common fish species (e.g., Dodd et al. 2006).  In other words,
the number of glochidia that encyst on a host has as much to do with the
host as it does glochidia viability in and of itself.  Somehow, there
would need to be the presence of a known host with the glochidia to
determine an answer to EPA’s proposal.  However, viable host species
are not known for many mussel species.  Therefore, this evaluation would
require testing with a few mussel species for which host species are
known with certainty.

George Dixon	I am not a mussel ecologist, but the approach taken, and
its implications for further research appear valid. Having said that, it
is likely that modifications to the approach will have to be made as
such research proceeds. It is also apparent that there is sufficient
species variability in the life habits of glochidia that a single
approach will not fit all species. 

Jerry Farris	The support of life history information and exposure
information can support this contention, but the pertinence of this
endpoint would continue to be challenged for its scientific
defensibility as relates to what we currently understand of exposure
durations related to this life stage and variety of adaptations.  

Bill Goodfellow	See comment for Question 3, the same comment is relevant
for Question 4.

Jim Pletl	As discussed earlier the uncertainties of using juvenile
mussel data for deriving water quality criteria are significant, but EPA
has identified additional concerns associated with tests of glochidia. 
The sensitivity of glochidia as a function of exposure duration, within
the constraints of the Guidelines, precludes the use of glochidia data
to derive water quality criteria.  This is primarily due to the
relatively short duration of instream exposure that can occur with this
lifestage and the sensitivity of glochidia to that duration of exposure.
 The issue of exposure duration instream versus that of lab toxicity
tests is critical to the use of all toxicity data in deriving water
quality criteria and is not unique to freshwater mussels and their
glochidia lifestage.  Decapod crustacean larvae are sensitive to
duration of test exposure without food for durations required by the
Guidelines.  The literature refers to the “Point of No Return” or
the PNR (Klaus Anger, various papers), where larvae will not survive
beyond a certain time period without food even if food is provided to
the larvae.  Toxicity tests of duration approaching the PNR are likely
stressing the test organisms in addition to the toxicant in question. 
The same can be said of glochidia tests with durations of 48 hours and
more and likely durations even greater than 24 hours.

Development of criteria must consider all test factors that might bias
test results in comparison to responses that are expected instream, and
exposure constitutes one of the most important factors.  Exposure is a
function of frequency, magnitude and duration.  The default frequency
assumption for water quality criteria toxicity tests is that exposure is
continuous.  Magnitude of exposure is usually dependant on
concentration, which may also be a function of other factors like pH,
alkalinity, DOC, etc.  The magnitude of exposure for glochidia will also
be affected by the species-specific conditions unique to glochidia of
different species.  For example, some glochidia are released into the
water column singly; some are released in groups encased in a gelatinous
shell where exposure would be mitigated.  Duration of test exposure
should be related to the duration of the lifestage, unless it can be
demonstrated that a different test duration is representative of that
lifestage.  This is a fundamental of the Guidelines although it is not
clearly delineated.  In the case of glochidia 48 hour or longer test
durations have been shown to result in responses not expected instream. 
This is a defensible reason to reconsider use of the acute glochidia
data to derive water quality criteria and defines well the research
needs for this taxonomic group and lifestage.  Shorter duration tests
will be necessary to provide reliable test results for deriving water
quality criteria. 

Hyalella azteca position statement and proposed rationale 

5.	Are the position statement and supporting rationale regarding use of
toxicity data for Hyalella azteca in criteria development reasonable and
scientifically sound recommendations?  

Alex Barron	Yes.  The EPA Guidelines for Diving Numerical National Water
Quality Criteria for the Protection of Aquatic Organisms and Their Uses,
in section IV. 5. H specifies that in order to be useful for calculating
a Final Acute Value, the must be relatively good agreement of the data
within a species.  Generally “if the acute values differ by more than
a factor of 10, some or all of the values probably should not be used in
the calculation” of a Species Mean Acute Value (SMAV).    This applies
to the available data set for Hyalella azteca, where the normalized
acute values listed in Table 2 range from 1.58 mg/L to 83.9 mg/L,
differing by as much as a factor of 53.  This wide range of acute values
for H. azteca required more careful review before determining what if
any of the data for this species is acceptable for criteria development.
  EPA has conducted an additional review of these data and has
determined that there are significant uncertainties with data for this
species, enough to preclude their use in deriving criteria.  The
position statement provides a summary of the information known
concerning possible factors that could cause the acute toxicity values
for H. azteca to be affected by some water quality characteristics not
previously recognized or considered and not fully understood at this
time.  Based on the current information available, it appears reasonable
that until these factors are better understood, the data for H. azteca
should not be used in the calculation of a SMAV or a FAV.

Steven Canton	I support the position statement based on the
workgroup’s review of a number of toxicity tests on Hyalella,
referenced in the rationale.  I can’t help but wonder if this
recommendation holds for Hyalella test data for other EPA criteria
documents, as well – for example, Hyalella is the most sensitive
species in EPA’s chronic cadmium database (EPA 2001).  Is this
position statement generic to all EPA criteria databases, or simply this
ammonia update?  

I should note that our prior review of the study also recommended
removal of Hyalella from the ammonia database, primarily because of the
poor control performance (AWWQRP, 2006), which it now appears may be due
to the ionic balance issue noted by the position statement.  So, I
support non-use of Hyalella toxicity data for this ammonia update.

Most of the recommendations provided have scientific basis. It appears
that ongoing studies will provide quality data that will widen the
knowledge for water quality requirements for H. azteca husbandry.
However, the statement concerning the complications due to possible H.
azteca genetic or taxonomic diversity is not relevant because the way
this information is used in criteria development, which only considers
appropriate data at genus level (i.e., GMAVs) not new species or
subspecies. In addition, the issue of a wide range of surface waters
inhabited by H. azteca could be easily addressed setting a site-specific
standard, which will consider only the species present at a particular
site.  Thus, inclusion of this information may add confusion rather than
help explain a decision on non-use for criteria development.



Jerome Diamond	I have reviewed the Hyalella data in a previous outside
peer review for EPA and agreed with EPA’s concerns regarding Hyalella
test data for ammonia.  I agree with the position statement in Appendix
B that Hyalella toxicity data should not be used in water quality
criteria development.  There is a clear need to resolve water quality
requirements and other methodological issues for this species in
water-only tests.

George Dixon	The exclusion of Hyalella azteca data, as summarized on
page 13, is fully defensible. I would suggest that while all of the
points are relevant, the pivotal consideration is the impact of chloride
and bromide ion on the viability of the species and the potential
interaction of the concentration of these ions with ammonia toxicity.
The Hyalella data will be suspect until that information is obtained.

Jerry Farris	Those points are very critical given the consideration for
osmoregulation by this organism and the points cited in the
workgroup’s review.  This seems well illustrated in the background
material furnished for this review.

Bill Goodfellow	I believe that the position statement and supporting
rationale is appropriate and justified.

Jim Pletl	The position statement is written to address the use of H.
azteca data for the ammonia update; however the issues presented would
likely be important for many other water quality criteria using data for
this species.  Therefore the position statement is not reasonable and
scientifically sound if H. azteca data are considered within the context
of all water quality criteria calculations.  There is a concern that
decisions such as this do not have associated rules as normally defined
in the Guidelines.  EPA is being reasonable and responsible in
considering factors that may bias the development of water quality
criteria in this case, but it is unclear whether EPA has a process to
identify such factors as new information is collected over time. 
Perhaps there was insufficient data to identify the H. azteca issues
when data for this species was considered for use, but EPA needs to
develop and implement a system to periodically review data for issues
such as this and publish the system as part of the Guidelines.  

Based on this comment the position statement may be reasonable but the
scientific basis of the recommendation is questionable without a well
defined system of review.  Such a system should include review of data
relative to standards of quantity and quality required to make such an
assessment.  EPA needs to determine how much information is necessary to
make this type of recommendation as well as the qualitative nature of
the data (measured vs nominal concentrations, static vs flow through
tests, etc.)  Without a system it will become increasingly difficult to
be consistent in making these types of decisions as more variances
arise.

  SEQ CHAPTER \h \r 1 Chronic Freshwater Criterion

6.	Are the toxicity tests and other studies used to derive the criterion
scientifically defensible for such use?  Are you aware of other relevant
data that were not used?

Alex Barron	Yes.  The toxicity studies referenced all appear to be good
quality studies and provide useful information on the chronic toxicity
of ammonia to aquatic organisms.  The use of EC20 values calculated by
regression analysis is excellent and provides a consistent estimation of
low-level effects under chronic conditions. 

All toxicity data relevant to ammonia toxicity that I am aware of are
included in the references and were considered in the reassessment
document.  

Steven Canton	While, the toxicity tests used to derive the criteria are
generally scientifically defensible (see discussion below); I am
uncomfortable with the calculation because there is insufficient data to
evaluate intra-species comparisons using the guidelines established for
deriving water quality criteria.  More specifically, the 2009 data set
still does not fully meet the EPA guidelines for developing water
quality criteria.  A salmonid representative has been added to the data
set (which was missing in the 1999 document), but a representative from
the Class Insecta is still missing.  It appears that the option for
using acute-to-chronic ratios for determining chronic criteria would be
more relevant than simply using a “hypothetical GMAV for insects”.

One of the studies noted in Question 1 may provide additional relevant
chronic data.

It is also important to point out that there are only two studies
(Anderson et al. 1978, Sparks and Sandusky 1981) available for Musculium
genus. Interestingly, those two studies provide substantially different
estimates of what level of ammonia is protective to fingernail clams –
creating uncertainty in the true chronic value.  A closer analysis of
these two papers reveals that while the later Sparks and Sandusky study
was designed to confirm Anderson et al’s findings, the researchers
faced some serious methodology deficiencies, as the authors recognize:
“Although it would have been desirable to perfect the culture methods
first and then employ them in the toxicity tests, methods were developed
as testing proceed because of limited time” (Sparks and Sandusky
1981). As part of their study design, Sparks and Sandusky used
clinoptilolite to remove ammonia, but note this compound could also have
remove potassium and chlorinated hydrocarbons – which raises the
question that toxic effects observed in Musculium transversum in this
particular study may be also attributed to other confounding factors in
the test water and not exclusively to ammonia exposure. In addition,
authors conclude that also biological processes (i.e., naturally
occurring bacteria that converts ammonia into a relatively non-toxic
nitrate) may be involved in ammonia removal rather than physico-chemical
processes. Thus, it would be advisable to re-analyze the relevance of
these studies for this update.  As it stands, I cannot fully support
calculating a GMAV for this fingernail clam using these two studies.

Another issue that is relevant to mention is the continued inclusion of
a temperature relationship for chronic ammonia criteria originally used
by EPA (1999), which was derived from a single study. It appears the EPA
used the Arthur et al. (1987) study, which evaluated acute ammonia
toxicity to 14 species (9 invertebrate and 5 fish species), to
incorporate temperature dependence into the chronic equations. There are
three problems with incorporating temperature in the chronic
relationship, but not acute. First, there is lack of chronic ammonia
toxicity studies for fish or invertebrate species to demonstrate the
relationship between temperature and chronic ammonia toxicity. Second,
the rationale for deriving an invertebrate chronic temperature slope
from acute data (i.e., fish and invertebrates) is unclear. Finally, the
assumption that chronically exposed invertebrates will have similar
temperature dependence compared to acutely exposed individuals is based
on the Thurston et al. (1984) study and those authors indicated that a
96-h period is insufficient to determine an acute toxic concentration of
ammonia for insects. Consequently, it appears there are no appropriate
data to support incorporation of a temperature component to the chronic
ammonia standards, just as EPA concluded with the acute equations. The
position statement notes that the temperature-dependent toxicity model
now only applies to invertebrates. 

I have to admit I was hoping the actual equations, inflection points,
and pH and temperature relationships would be thoroughly re-evaluated as
part of this update, given the abundance of new data included not in the
1999 document.  I would strongly recommend re-evaluation of the chronic
temperature component of the equations rather than just accept the 1999
update evaluation.  Although, I have to admit, the draft document
doesn’t actually provide any equations to test – just recalculations
of acute and chronic endpoints.  This is another disappointment.

Jerome Diamond	Most of the chronic toxicity tests relied on by EPA for
chronic freshwater criterion development are consistent with EPA’s
Guidelines and other water quality criteria developed.  I am not
convinced that the 28 day juvenile test data for the snails included in
Table C should be used.  Both of these tests used mixed-aged organisms,
and in the case of the Ozark spring snail, organisms were also
field-collected adding uncertainty in terms of their condition and
acclimation to laboratory test conditions.  In general, EPA has not
relied on such tests for chronic criteria development, and I believe for
good reason.  I am not aware of any other relevant data that EPA did not
already consider in this document.

George Dixon	Please see the response to question 1 above. That statement
is also applicable to the chronic data base.

Jerry Farris	The tests utilized for the criterion are scientifically
defensible for such use and I’m not aware of other relevant data that
were not used.  

Bill Goodfellow	I believe that they are scientifically defensible for
the use in the development of the chronic criterion. I am not aware of
other relevant data.

Jim Pletl	Using the standards established by the Guidelines and the
respective toxicity test methods the toxicity tests used to derive the
chronic criterion are scientifically defensible.  However, this does not
mean that the data generated from these tests and the resulting test
endpoints are scientifically defensible for the use.  As stated before,
there are standards that should be met to establish that the test
results are of sufficient quality to be used in deriving water quality
criteria.  These standards include an appropriate concentration response
curve and limits on variability within and between test treatments,
replication of tests and use of reference toxicant tests to gauge
organism sensitivity, as well as defensible control response across
tests.  One of the shortcomings of the chronic freshwater mussel method
is that it does not include minimum organism size (weight, length)
requirements that must be met at the end of the test in controls.  It is
unknown whether the growth expressed in the controls of these tests is
acceptable in a natural environment where stress does not exist.  The
tests only can establish weights attainable in controls in the lab
environment when organisms in the controls meet the minimum survival
requirement.  The organisms may be stressed in all test vessels,
resulting in a bias of the test endpoint.  The combination of treatment
stressors (ammonia and vessels without sediment, for example) may
produce a synergistic effect resulting in a lower ammonia IC20 for
growth and survival than that resulting from a test which has only one
stressor (ammonia).  In this case, the control response cannot account
for the synergism, even though the control is intended to account for
all factors independent of the tested treatment.

I am not aware of any other relevant data that was not used.

7.	Is the freshwater chronic criterion scientifically defensible with
mussels present and mussels absent?  

Alex Barron	Yes.  Similar distinctions for the presence or absence of
certain sensitive organisms showing different sensitivities to ammonia
have been used before.  The 1999 ammonia criteria recommended different
chronic criteria based on the presence or absence of early life stages
of fish in the water body, setting a precedent of allowing different
criteria for different situations.  These different criteria were based
on a similar magnitude of difference in sensitivity between the early
life stages and the older life stages of fish (in the 1999 reassessment
criteria) as is seen between the sensitivities of freshwater mussels and
other species (in the 2009 reassessment).  This policy of providing
these separate criteria based on the presence absence of freshwater
mussels will provide some flexibility in implementing the criteria in
some instances and is reasonable.  

Given the wide spread nature of ammonia in natural environments and the
difficulties and expense imposed on municipalities in treating for
ammonia in sewage treatment facilities, a significant lowering of the
ammonia criteria is likely to have important consequences for
municipalities.  Providing these two criteria may help alleviate
concerns that any additional restrictions on ammonia discharges are
targeted to providing the needed level of protection for the waterbody,
based on whether or not the more sensitive mussels are present.  
Providing two EPA recommended criteria that already take these issues
into account could help reduce the potential for requests for
consideration of developing site-specific criteria based on the
recalculation procedure, which can require significant resources to
pursue. 

Table C shows the four most sensitive genera and the calculation of the
CCC for the two different criteria (with and without Unionid mussel
data).  The data under “excluding freshwater mussel data (family
Uniondale)” includes the GMCV for the fingernail clam, Musculium
transversum as the third most sensitive genera and this GMCV is used in
the calculation of the CCC for “non-Unionid containing waters”. 
While it is true that M. transversum species is not in the family
Unionidae, it is a bivalve mollusk.  

In the recommendations for developing separate acute criteria, the
calculation of the FAV and CMC as shown in Table B, the Asiatic clam was
removed from the dataset to produce a criteria based on a “non-bivalve
mollusk” dataset (as commented on under question # 2 above).   There
may be water bodies with physical habitats (such as some headwater
streams) that are unsuitable for bivalve mollusks of any kind, not just
Unionids.   EPA should consider whether or not to base their
recommendations regarding separate criteria based on removing data for
Unionid genera only, or for all bivalve mollusks.  The same approach
should be used for both the acute and chronic criteria.   

(Note: there is a small typo in Table C; the far right hand column
should be labeled GMCV, rather than GMAV)

Steven Canton	The freshwater chronic criterion based on a mussels
present/absent approach maybe scientifically defensible – but will
potentially result in very different approaches, even in regions of the
United States where mussels are known to be absent or rare.  In fact,
the primary question will become “what level of data is required” to
determine whether the habitat is or is not suitable to support mussels
and/or whether mussels have ever occurred within the site historically. 
Guidance on such data needs questions would be useful to include in the
criteria document to support a with/without mussel criterion.  

Additionally, even when the “mussel present” option is applied, a
whole other series of issues arise, because there is often a poor
correlation between the water column NH3 concentration and the pore
water NH3 concentration where most juvenile mussels reside. 
Site-specific sediment conditions can greatly influence the availability
of NH3, and in the case of the upper Mississippi River, as much as 6-30
times greater than the surface waters (Bartsch et al. 2003). 

In such cases, criteria developed without considering the complete
affects of sediment on the accumulation of NH3, or other confounding
affects on mussel communities (i.e., the fact that mussel beds can
self-produce nutrients, such as ammonia) (Dankers & Koelemaij 1989), may
greatly underestimate, or even overestimate, a criterion that is
protective of mussels.  In addition, this also brings up the question of
whether the use of sediment is appropriate in chronic toxicity tests
designed to be protective of aquatic life in the water column, and what
endpoint should be evaluated using such methodologies.  

Jerome Diamond	I don’t think it is.  See answer to question 2.  The
chronic value used for L. fasciola in Table 4 (0.39 mg N/L), and which
must have been used in the GMCV calculation in Table C, doesn’t appear
to make sense given an IC25 at the same pH and temperature reported of
0.38 mg N/L on p. 20 and the IC25 based on the actual data (pH = 8.2 and
20 ( C) of 0.39 mg N/L.  The IC or EC20 should be lower than the IC or
EC25.  The IC pin value of 0.23 mg N/L appears more in line with the
data for this species.  EPA should at least defend why 0.39 mg N/L is a
more appropriate value than 0.23 mg N/L for this species.  In addition,
wouldn’t it be more consistent to use the ICp value for both Lampsilis
species because the L. siliquoidea chronic value could only be
calculated using the ICpin method?

George Dixon	The chronic criterion is defensible, provided that suitable
guidelines are developed to differentiate environments where mussels are
present from those where they are absent. Please see my response to
question 2 above. 

Jerry Farris	The criterion does not seem defensible for locations where
mussels are present as would relate to considerations under the
Endangered Species Act, but would likely be defensible under
considerations with the Clean Water Act.   I pose this statement as an
example of the difficulty in relying on scientific defensibility in the
face of a very broad range of considerations in risk and decision
management that may rely on these estimates.  

Bill Goodfellow	Similar to the acute criterion, I believe it is
scientifically defensible to develop the criterion with mussels present
and mussels absent. I do believe that the EPA should have a cautionary
note as in the acute criterion as to when and how to determine if
mussels are present or absent.

Jim Pletl	As stated in question #2, EPA should determine whether simply
the presence of mussels is important or the presence of specific species
is important.  The first step in determining whether mussels are present
or absent is to define a standard for the conditions describing that a
population exists at a site.  This is not a straight-forward exercise
and can be quite difficult as seen in various site-specific criteria
projects around the country.  Guidance must be provided as to what
parameters and measures of those parameters are to be used to define
presence/absence of a population.  Quantitative evidence of reproduction
would be required at a minimum, as well as the presence of various
developmental stages.  The normal abiotic conditions documented for a
species should also be compared to that of a particular surface water to
determine whether a conclusion of presence or absence is supported. 
Certainly the availability of host species could influence the
determination of absence/presence.  

Once absence/presence is established EPA should only use mussel data to
develop water quality criteria for waters where mussels are deemed
present.  EPA should ensure that conditions of the site where absence
has been concluded have not been anthropogenically changed to a state
where mussels cannot reproduce and biological records have documented
the presence of previous populations.  Historic records of water quality
and biological surveys would help support this determination.  It is
clear that EPA and the states will share the burden of the
presence/absence determination and that there will be differences of
opinion as to whether, when absence seems apparent, conditions changed
due to human activity that could have been reasonably precluded or the
species simply never successfully established a population at the site
due to species-specific factors.  Given that the chronic sensitivity of
mussel species drives the CCC approximately 44% lower than when mussel
data is not used the management actions and costs based on criteria
without mussel data could be significantly different than when mussel
data is used to develop the CCC.

The presence/absence of species is usually a question that is asked when
criteria are recalculated or a site-specific criterion is being
developed.  In these cases the issue of removing species from a database
is addressed by determining whether data for other species with similar
taxonomy is available.  The taxonomy and reproductive cycle of the
freshwater mussels is so unique that this will not possible.  Therefore
a decision must be made as to whether to apply a CCC with or without
mussel data.  Given that there is so little chronic data available for
these mussels and that there are outstanding issues that must still be
addressed (use of shell measurements to represent growth, tests without
sediment, irregular concentration response curves, etc.) a CCC with
mussel data will result in water quality standards with uncertain
consequences.  This uncertainty demands that criteria without mussel
data be available for states to use when mussels are not resident
(reproducing population found year to year).  At this point in time a
CCC without mussel data would be more scientifically defensible than one
with mussel data.

Use of 28-day Juvenile Test Data 

8.	Given that the juvenile life stage of freshwater mussels is
relatively long (2-6 years) are 28-day exposure tests with juvenile
mussels scientifically defensible as "chronic" test data for criteria
development?   

Alex Barron	I do not believe that growth measured during a 28-day test
can be considered a true measure of a chronic effect as intended by the
requirements in the 1985 Guidelines for Aquatic Life Criteria, such that
they can be related to significant population impacts. Given the life
span of the organism and the time needed to reach maturity, a 20%
difference in growth observed over 28 days may be reversible and may not
be relevant on a population level and should not be used to derive
criteria. 

Steven Canton	Water quality criteria for the protection of aquatic life
are derived using toxicity endpoints that relate to population level
impacts.  In general, these endpoints relate to survival, growth and/or
reproduction.  I would agree that the 28-day test with juvenile mussels,
while similar in duration to a standard chronic test for many other
groups of organisms, is not technically an early-life stage test
according to the 1985 Guidelines for Aquatic Life Criteria, and as such,
should perhaps not be included in the chronic database.

If this is carried forward, I would support the approach of the 28-day
exposure tests being considered as “other data’ – if the test are
acceptable in all other respects.  Reasons for not including in chronic
criteria calculations could follow and build on the arguments presented
in the position paper:  

First, it is not representative.  If the duration of the life stage is
2-6 years for juvenile life stage of fresh mussels, 10% of this amount
of time would be 73-219 days. Thus, a longer exposure (i.e., 90-day)
would be more representative for a “chronic” study for juvenile
fresh water mussels than 28-day tests. 

Second, extending the exposure time would allow the tested animals to
adapt better to laboratory conditions, thus reducing the stress and
decrease the risk of having test condition-biased results. 

Third, although it will be more challenging to keep all the water
quality parameters under control in a longer exposure it could also
provide other avenues for future research on juvenile mussels (i.e., if
captivity conditions and pollutant exposure would enhance growth over
longer periods of time, if they may need additional food source, among
others)

Once possible conflicting issue is that other species used in EPA
criteria development have “allowable” chronic test conditions that
also represent a small portion of their total life span – e.g., 60-day
post hatch ELS for salmonids, which can live for years.  The difference,
conceptually, is that the 28-day test for mussels does not include the
earliest life stage prior to juvenile testing commences – i.e., it
does not start with egg, then hatch, then juvenile exposure more typical
of an ELS for fish.  If we can conceptually overcome this “weakness”
for a 28-day juvenile mussel test, then I suppose growth-effects data in
a 28-day juvenile mussel test could be at least evaluated in the context
of other reported data – and if comparable to effects data from
longer-duration testing, then perhaps it could be used in some form for
criteria calculations.

Jerome Diamond	The question posed here by EPA implies that a 28-d test
may not be sufficiently long for freshwater mussels that live 2-6 years
(actually, many unionid species live far longer than that—20 years is
not uncommon).  If this is the case, I fail to see how EPA justifies use
of 28-d tests for fish such as salmonids that live 3-5 years or sturgeon
that live > 30 years.  I realize there’s McKim’s 1977 paper but that
was based on outdated test methods and has questionable relevance to
data collected in more recent years.

The discussion concerning the usability of 28-d mollusk survival and
growth tests in chronic criteria development (pages 15-16 of the draft
document) appears to have contradicting statements and questionable
rationale in my view.  The discussion begins with the statement that
28-d tests do not qualify as life-cycle, partial life cycle, or early
life stage tests as defined in EPA’s Guidelines.  According to those
guidelines, 28-d tests can only be used for fish because it has not been
demonstrated that a 28-d exposure is a reasonable predictor of
invertebrate sublethal chronic effects (defined as at least 90d for
vertebrates and presumably for invertebrates such as mussels which live
for several years).  However, the draft document goes on to say that
28-d survival information could be used to inform the chronic criterion
because obviously, lethality is not reversible.  Yet in footnote #2 on
page 15 of the document, as well as in Appendix C, uncertainties
regarding the 28-d mussel test method are discussed which include
optimal quality and quantity of food, exposure apparatus, etc. needed to
sustain juvenile mussels in good health and maintain adequate growth
throughout the test.  Based on my limited experience testing bivalves, I
believe it may not be as challenging as one might expect to keep control
juvenile mussels alive for 28-d given adequate oxygen and clean water
but it may be quite another thing to document that organism condition
(or alternatively, susceptibility to stress) is truly adequate. 
Therefore, the fact that juvenile mussel survival is lower with exposure
to certain levels of ammonia may be due to having an already stressed
population due to inadequate food or other factors.

A similar finding is now being made by EPA after reviewing chronic
Hyalella chronic test data (reported to be one of the most sensitive
species in the 1999 ammonia criteria document).  In that case, control
survival was apparently satisfactory but test organisms were in fact
ultrasensitive to the chemical due to suboptimal test conditions.

A second point I would raise regarding the chronic mussel tests is that
they were only conducted once (based on published information) for a
given species in one laboratory.  While this in itself is not a reason
to exclude test data, it is clear that there is considerable variability
in acute endpoints among juvenile mussel tests using the same species
and conditions (even within the same lab as mentioned previously), which
should apply to chronic survival with ammonia as well.  Given that
certain mollusks appear to be far more sensitive to ammonia than any
species tested thus far, there should be some replication of these
results to be sure we’re not going to find out that EPA needs to do
yet another reassessment because of issues similar to the ones they are
dealing with now for Hyalella.

It is important to note that while interlaboratory testing was conducted
for acute juvenile tests (Wang et al. 2007), similar interlab testing
has not been conducted (or at least published) to my knowledge for 28-d
mussel chronic tests.  In addition, one should view the acute
interlaboratory results cautiously because, as mentioned previously, a
single culture and water source was used by all laboratories in that
study.  As Wang et al. (2007) point out:  “However, the present study
used the same batch of test organisms and the same dilution water, which
were different from the referenced inter-laboratory studies and might
have contributed to the lower variability in test results. The present
study was designed to determine the inherent variability in the test.
Higher variability would be likely if the inter-laboratory tests were
conducted with different sources of dilution water and organisms (e.g.,
from different populations or watersheds). Therefore, additional study
is needed to further characterize potential variability associated with
the newly developed ASTM standard methods for conducting acute toxicity
tests with early life stages of freshwater mussels.”  Therefore, the
CVs reported by Wang et al. (2007) are likely to be underestimates of
true test method variability.  This is borne out by the acute data
generated for the same species by different labs as noted previously in
my comments.

George Dixon	This is well covered in Appendix C, and the arguments are
convincing. I agree that the 28-d test does not qualify as a chronic
test, since it is neither a life-cycle test nor a partial life-cycle
test. I also agree that a concentration causing a greater than 20%
reduction in survival could be used as an upper limit for the Species
Mean Chronic Value, assuming the test met the requirement for good lab
practice. 

Jerry Farris	Research that validates effective endpoints related to the
range of pollutant effects in freshwater mussels have shown that 28-30
-day exposure tests in both field and laboratory settings can be
included to predict long term impacts related to biomarker responses,
bioaccumulation, and even comparisons to conventional “chronic” test
data.   The 28-day test exposure period will continue to be temporal as
relates to the life cycle, but can be relational to long-term
considerations and offer critical information when combined with
multiple lines of evidence.  Inclusion of more than one life stage
response also advances mussel toxicity assessments, rather than
providing semantic arguments over longevity for a life stage and whether
the exposure period qualifies as “chronic”.   Data from the 28-day
test can be defensible when supported by “other” data.  Other data
should also be scientifically defensible.    

Bill Goodfellow	I do not believe that 28 day mussel studies should be
considered as chronic test data for the national database as described
in the position statement and supporting documentation. I believe this
is very well thought out and applaud the developers.

Jim Pletl	This question is not only a function of the duration of
exposure in the test but whether the test meets the requirements of the
Guidelines.  Clearly the test does not meet the Guidelines requirements
because it does not address either the entire life cycle or, for partial
life-cycle tests, the appropriate life stages and test endpoints defined
by the Guidelines.  The reliability of this test is also clouded by the
apparent impact of testing juveniles in the absence of sediment and the
use of shell length as a chronic test endpoint for freshwater mussels
(see comments below).  Perhaps the Guidelines require updating (has not
been formally revised since its first release in 1985), but until this
is done the 28-day juvenile test does not meet the standard for
scientific defensibility that EPA has established.

One might suggest using the data from these tests as an estimate of
freshwater mussel chronic toxicity because the tests did not include the
glochidia life stage and it is more appropriate than not using the data
for criteria development.  Unfortunately there is enough concern
regarding the reliability of this test even to accurately represent the
sensitivity of juveniles that it would not be appropriate to use the
data in any water quality criteria capacity.

Independent of the issues raised above, it would seem that a 28-day
exposure test with juvenile mussels would be representative of the
sensitivity of juveniles.  However, chronic tests used for developing a
CCC are intended to evaluate or extrapolate the sensitivity of a
population to a stressor.  This requires reliable knowledge of the
sensitivity of the most sensitive life stages of that population.  It is
unknown how the sensitivity of glochidia compares to juveniles because
of the issues associated with both acute and chronic test methods. 
Without this knowledge and reasonable certainty it cannot be determined
that the 28-day juvenile test with freshwater mussels is scientifically
defensible relative to development of a CCC.  

9.	Should toxicity studies on the growth rates of mussel shells during
28-day tests be considered quantitatively when developing water quality
criteria?

Alex Barron	No.  I do not believe that differences in growth rates
measured over a 28-day period should be considered to provide acceptable
quantitative chronic endpoint data for freshwater mussels that should be
used to calculate chronic criteria for ammonia. Chronic criteria for
ammonia must be based on clearly recognized adverse effects in order to
be acceptable.  Reductions in survival are irreversible and are clearly
adverse effects, while slight reductions in growth may not be
significant to the population, or the organisms may be able to
compensate over time and reverse the difference in growth.   These type
tests do not meet the requirements of the EPA Guidelines for use as a
quantitative value in driving a chronic criterion.   Using such data to
derive chronic criteria, especially if it significantly influences the
final criteria would likely increase uncertainty about the final
criteria. 

Steven Canton	As described in the previous question, the studies on
growth rates of mussels shells during 28-day tests should only be used
as information to better evaluate appropriate longer-term chronic
studies (i.e., 90-day studies). The results from longer chronic tests
could certainly be used quantitatively for the developing the water
quality criteria. Unless 28-day test data provide similar endpoints to
long-term testing, it would be necessary to keep the results separate
and potentially have two conflicting sets of data (i.e., 28-day and
90-day data) – which will only make the calculations of the new
criteria more complicated, and can lead to confusion when trying to
implement.  And I have to say that ammonia criteria development is
confusing enough as is.

Jerome Diamond	I don’t believe there is sufficient information or
adequate test method development and standardization at this time to be
able to use growth rates of mussel shells in 28-d tests for criteria
development.  Growth rates, based on shell measurements, are not
sufficiently large over a 28-d period to have much confidence in the
results.  Based on the studies reported in this draft, as well as others
of which I’m aware, shell length is too coarse a measurement to be
used as an indicator of mussel growth over such a short time period. 
Furthermore, it is generally not known what the expected growth rate
should be for a given species under these test conditions.  As I
indicated earlier, without having such information, I don’t believe
that survival data from such tests can be used with confidence as well.

George Dixon	The approach taken in the document (Appendix C), that of
not using data for growth of organisms in non-life-cycle tests with
durations of less than 90 days in the derivation of a Species Mean
Chronic Value (SMCV), is in my opinion perfectly appropriate. The
justification given in paragraph 5 of Appendix C centers on two points:
lack of direct comparisons of growth between shorter tests and
life-cycle tests and uncertainties around the relationship between
short-term growth reductions and growth effects over a full life cycle.
Both are valid reasons to exclude the data from calculations of SMCV.



Jerry Farris	Growth rates of mussels during the 28-day tests should not
be considered quantitatively until more information is available on the
validity of this test.  Variability is still attributable to diet and
holding conditions and has been cited broadly.  

Bill Goodfellow	The problem with using shell growth rates, it doesn’t
necessarily link to a smaller organism being any less viable or having
less reproductive benefits.  Since mussels are not a commercially or
recreational important organism, I believe that the use of growth rates
are not valid as compared to a fish species or commercially important
bivalve.

Jim Pletl	Review of Wang et al. (2007) for both the copper and ammonia
28-day tests shows that there is a very small difference between the
length of the shells at the beginning and at the end of the test based
on control responses.  In fact it appears that some of the tests show no
statistical difference in shell length between the beginning and end of
a test at alpha=0.05 and assuming a normal distribution.  This does not
support the use of shell length to represent growth in the 28-day mussel
tests.  Further this paper shows that using shell length to represent
growth does not provide reliable dose response curves in most cases. 
Although the authors only conducted 21-day tests both Bringolf et al.
(2007a) and Bringolf et al. (2007b) found that the concentration
response curves for different chemicals were not what one would expect
for a toxicant.  Although there may be issues with organisms benefitting
from certain concentrations of chemicals, this fact only complicates the
interpretation of this data and increases the uncertainty of using
toxicity test results based on shell length.

10.	 Regarding the position statement and rationale on use of juvenile
mussel growth data – 

Is it scientifically defensible to include the juvenile growth data from
a 28-day exposure period as “other data” that might influence the
criteria however not be used directly in the derivation of the criteria
value? 

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

Alex Barron	Is it scientifically defensible to include the juvenile
growth data from a 28-day exposure period as “other data” that might
influence the criteria however not be used directly in the derivation of
the criteria value? 

Yes.   Although too many questions surround the use of growth effects
during a 28 day test in relation to long lived species such as
freshwater mussels to consider using differences in growth  as data that
directly influence the calculation of the final chronic criteria, these
data still deserve consideration and inclusion in the EPA criteria
document.  These data do provide useful information to help judge the
potential for “other” effects that have a potential for concern.
Including such data under the category of “other data” can provide
additional information that can lend confidence that the criteria based
on survival can provide an adequate amount of protection.  The magnitude
of difference in growth may make a difference; a 20 to 25 % reduction in
growth may not be of much ecological significance, especially if there
is a thought that the effect may be short term and reversible, while a >
75% reduction in growth may be considered more of a concern.  The
magnitude of difference between the effects on survival and the effect
on growth can also provide useful information.  Altogether, I see value
in presenting these growth data under the classification of “other
data” in the criteria document.

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

I would agree that this statement could also be true.  We simply do not
know whether or not an effect on growth observed over a 28-day exposure
can be considered reversible or indicative of a permanent effect.  As
discussed above, I believe the magnitude in difference in the growth and
the magnitude of difference between the survival end point verses the
growth endpoint could influence the level of concern regarding such
data.  

Steven Canton	Is it scientifically defensible to include the juvenile
growth data from a 28-day exposure period as “other data” that might
influence the criteria however not be used directly in the derivation of
the criteria value? 

My understanding of data included as “other data” in the criteria
document is that these data are provided just as a reference to make the
interested parties know that these data exist.  They can certainly be
presented that way and not considered in the calculations of the chronic
criteria for the reasons outlined in the previous two questions. 

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

As noted in the position paper, it is not possible with the available
data to conclude that impaired growth after 28-day exposure will
necessarily lead to continued reductions in growth over time.  The
position paper rightly points out the possibility of organisms to
“recover” from potentially short term stress and show no further
decline. In fact, these organisms may exhibit normal growth during a
more relevant long-term test.  However, given the lack of information
available to compare short-term and long-term toxicity studies in fresh
water mussels, it would be advisable to note, as appropriate, that any
of the three outcomes: continued decline, stabilize, or recovery is
possible. In addition, the propensity for any of these three outcomes
will most likely depend upon the particular mussel species and the stage
of development used in the toxicity tests.

Jerome Diamond	I agree with EPA’s position that mussel growth data
should be categorized as “other data” until it is demonstrated that
there is a repeatable, defensible methodology for generating adequate
growth data and that the growth endpoint calculated is ecologically
meaningful.  For this reason, I don’t think it matters whether
impaired growth observed in a 28-d test is indicative of effects at
longer exposures or not.  However, I would recommend that EPA not
consider 28-d mussel growth as a valid endpoint based on methodological
and perhaps biological uncertainties, as I mentioned under (9) above,
rather than because reduced growth rate is theoretically a reversible
effect.  I see no reason why reduced growth would not be an even bigger
effect with increased exposure duration to a toxicant.

George Dixon	With respect to the first point above, inclusion of the
data as “other data” is perfectly appropriate. This is consistent
with the practice for short-term growth results with other species, and
I see no reason why mussel data should be treated differently. With
respect to the second question, this is a statistically valid point, but
it does not reduce our level of uncertainty around the actual meaning of
the data. The issue will only be resolved when someone undertakes 90-day
tests. 

Jerry Farris	Is it scientifically defensible to include the juvenile
growth data from a 28-day exposure period as “other data” that might
influence the criteria however not be used directly in the derivation of
the criteria value? 

Growth should not be considered as defensible for the above mentioned
reasons relating to lack of reliable test conditions, but can be used as
supportive information that might influence the criteria.  Studies that
include recovery and or stability in growth or even degrowth with
freshwater bivalves, have produced more reliable estimates of pollutant
effects and could therefore be used to influence the criteria when
included with broader scale studies.  

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

As stated, the above consideration seems very confusing and should be
clarified to relate to the defensibility as considered with “other
data”.

Bill Goodfellow	Is it scientifically defensible to include the juvenile
growth data from a 28-day exposure period as “other data” that might
influence the criteria however not be used directly in the derivation of
the criteria value? 

I believe that it should remain in the other data section so the
relevant data can be used in site specific evaluations.

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

I don’t believe that data supports this and such a statement should
not be made. 

Jim Pletl	Is it scientifically defensible to include the juvenile growth
data from a 28-day exposure period as “other data” that might
influence the criteria however not be used directly in the derivation of
the criteria value? 

Based on the uncertainty associated with testing juveniles without
sediment and the lack of reliability in the shell length endpoint it is
not appropriate to use the chronic data in any way associated with
development of a CCC. 

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

The comments provided above on the reliability of the shell length
endpoint for chronic tests of freshwater mussels indicate that it is not
a reliable endpoint to represent the chronic sensitivity of these
organisms.  Without more data, it is not defensible to speculate on the
outcome of tests with longer duration when the reliability of the
endpoint being used is in question.

11.	The values of the acute and chronic ammonia criteria have a strong
dependence on pH.  Juvenile and adult mussels, as sediment-dwelling
organisms, inhabit a medium that may have vertical pH gradients, thereby
creating some uncertainty about the appropriate pH to assign as their
exposure conditions.  For applying a criterion protecting mussels, do
you have suggestions on how states and EPA might determine the pH
applicable to the sediment micro-environment to which mussels are
typically exposed?

Alex Barron	The toxicity data currently available for freshwater mussels
are for glochidia and juveniles.  The exposure to unencysted glochidia
will be primarily via water exposure, and once encysted on a host fish
they will be in a less exposed condition.  Juvenile mussels will also be
initially exposed to ammonia in the water column following dropping off
(excystment) from the host fish.  As the juvenile mussels are deposited
onto the substrate they may remain on the surface or may become situated
within a few centimeters of the surface. They will need to maintain a
position on or near the surface in order to ensure access to
well-oxygenated water as well as filterable food.  In all cases, the
exposure to ammonia will be over the gills via water; either surface
water or interstitial water and the water quality criteria can be
adjusted to account for any differences in pH between the water column,
the surface-sediment interface or pore water at a specific depth in the
sediment.  

For normal use in establishing permit limits, or assessment of
waterbodies; the criteria should be applied to the water column.  If
there is a special concern at a site, the option of monitoring the pH of
interstitial water can be investigated, and the pH characteristic of the
interstitial water can be used to assess the criteria.  Care should be
taken to determine the pH in the interstitial water at depths where
mussels are likely to be found at the site.

Steven Canton	Given that juvenile mussels of most species reside
completely in the sediment, filter pore water, and feed on the sediment,
I would suggest one approach might be determining a mean burial depth
for the juveniles of the selected mussel species within the 2009 data
set.  In laboratory tests, juvenile Villosa iris are reported to bury
<1cm, yet were not exposed to the overlying water (Yeager et al. 1994). 
Perhaps it would be useful to use this burial depth and evaluate the
average pH conditions within a burial zone, per se.  In addition, it is
likely any measures of pH in the mussel micro-environment (i.e., taking
measures of water quality over a mussel bed) may also show elevated
ammonia levels simply from the metabolic activity of those organisms,
themselves.  I can’t help but wonder how to determine attainment of
“safe” ammonia levels in future field monitoring, given the
potential for organism-generated ammonia values.

Jerome Diamond	I am not so sure that applying ammonia criteria to
juvenile mussels at least, and possibly adults as well, should consider
sediment pH at all for several reasons.  First, larger juvenile and
adult mussels are reported to be exposed to waterborne contaminants via
siphoning the overlying water and not as much via sediment ingestion or
pore water exposure during their growth season (Strayer et al. 2005). 
This is particularly true during times when water temperature exceeds
50( F (Watters et al. 2001; Schwal and Pusch 2007).  In fact, many
mussel sampling protocols specify a minimum water temperature of 50-60(
F because mussels are near or on the surface under those conditions and
more readily seen (e.g., Pennsylvania DEP and Army Corps sampling
protocol in the Allegheny River, PA).  During this time (typically
spring, summer, and fall in most waterbodies of the temperate U.S.),
mussels are actively feeding via siphoning the water column and sediment
pH is probably not a factor affecting water column ammonia toxicity. 
Rather it is the overlying water pH that is probably more important.

Second, many researchers report that sediment pH and ammonia can vary
substantially from one location to another within even small areas and
may be dependent on very complex redox gradients that are difficult to
measure much less predict (Burton 1992; Stemmer et al. 1990; Sarda and
Burton 1995).  It is no easy matter to identify an appropriate sediment
pH that should be used to apply an ambient water ammonia criterion.

Third, many of the mussel species tested thus far, or that are species
of concern, inhabit well-aerated sediments that often have either fairly
coarse particle size (gravels – cobbles) or sand, both of which have
interstitial water closely resembling the water quality conditions of
the overlying water (Salmon and Green 1983; Neves and Widlak 1988; Way
et al. 1990).  In these cases one would not expect a large difference
between sediment and overlying water pH.

Fourth, it seems premature to consider complex pH adjustments based on
sediments when it hasn’t been demonstrated that juvenile or adult
mussels are as sensitive to sediment ammonia as they appear to be to
water column ammonia in lab tests.  If indeed sediment pH needs to be
considered, then it seems likely that mussels are not being exposed to
surface water (including surface water ammonia) but rather sediment and
interstitial ammonia.  If this is the case, there is no need to adjust
water column ammonia criteria based on sediment pH.

Finally, basing a water column criterion on sediment water
characteristics represents a major departure for EPA in terms of
criteria development.  While I would encourage the Agency to consider
more holistic criteria that take into account multiple media (e.g.,
sediment and water), I don’t believe there is sufficient information
at this time to warrant altering a water quality criterion, which was
derived using water exposure test methods exclusively, due to sediment
characteristics.

George Dixon	While mussels live in sediment, the pH of importance from
an ammonia criteria perspective is the pH of the respiratory
environment, which in this case is the pH at the sediment-water
interface. The sediment pH is of importance only to the extent that it
influences the pH of the water at the interface. Remember these are
water quality criteria, not sediment quality criteria. I have no
suggestion as to the appropriate method for determining the pH at the
interface. The methodology, more in the realm of environmental
chemistry, falls outside of my area of expertise. 

Jerry Farris	Even if an adequate amount of information was available for
any benthic invertebrates for chronic toxicity related more specifically
to the generic relationship of ammonia toxicity to pH, larger issues
challenge this consideration with the reassessment for changes related
to pH shifts, which may increase when phase waters are isolated for
testing.  These relationships have only recently been examined for
juvenile freshwater mussels, without sufficient detail furnished as to
the pH adjustment used with diluter systems in those tests or their
effect upon alkalinity changes measured during the assays (Wang et al.,
2007).  Although these tests seem suitably performed to establish the
influence of pH on the acute toxicity of ammonia and even the pertinence
of normalizing ammonia toxicity data for mussels to a common pH end
point (total ammonia nitrogen at pH 8.0), information is still lacking
for the sediment relations specific to mussel responses.   As relates to
this consideration, the bulk of the studies that have examined sediment
and phase water relationships to date rely heavily on marine field
monitoring, and not standardized laboratory testing or field monitoring
with freshwaters (Salazar and Salazar, 2007).  Any scientifically
defensible way to account for these differences would greatly benefit
from studies with comparable methods examining habitat preferences,
feeding, and mobility of a sensitive mussel life stage (juvenile) among
sediment compartments, since the synthesis of ammonia toxicity data
confirmed its status as a sensitive species.  Data from these studies
would also be needed to generate pertinent modeling of the ammonia
fluxes from sediment and the concentration profiles in the phases that
would have consequence to mussel impact.   

Bill Goodfellow	It is much easier to do this for water column organisms.
I believe that the agency will open themselves up for tougher battle if
they attempt to make this issue even more complex than it already is. I
believe that noting that this is the case and stating that pH dependence
vs. the criteria as it relates to mussels was not developed at this time
given the uncertainty about what the appropriate pH assignment should
be.  

I also argue whether pH in sediment is a relevant question with regards
to criteria since the management tool for water quality is going to be
primarily the NPDES permit. How will anyone be able to control sediment
pH to manage it?  I am not sure what this does for development of a
defensible criterion.

Jim Pletl	It will be necessary for states and EPA to directly measure
the pH of the environment where each life stage exists.  For juvenile
and adult mussels this will require measurements of the sediment in each
location where the criteria are applied as water quality standards. 
Guidance from EPA on how to appropriately represent the pH of sediments
where ammonia criteria are applied to protect freshwater mussels will be
necessary.  Measurements in the field will also be necessary because the
pH of a discharge regulated by a NPDES permit cannot be used to estimate
the pH of the sediment.  The pH of the sediment can only be assessed
through direct measurement.  This may be done in situ or by rapidly
measuring the pH of relevant sediments collected with sediment grab or
core equipment.

12.	In general, should the criteria include a consideration for the
potential pH difference between sediment and the water?  If so, what is
the most scientifically defensible way to account for these differences
when deriving protective water quality criteria?

Alex Barron	The potential differences in pH between water and sediment
is not so much directly related to the criteria development, as it is an
issue related to how and where the criteria may be applied.  The water
quality criteria should provide recommendations for protective criteria
based on ammonia concentrations in water and adjusted for pH and
temperature.  On a site-specific basis, if in sediment there are
significant differences between pH in the interstitial pore water in
sediment and the overlying surface water,  this can be assessed by
site-specific measurement of pH levels in the sediment (at depths where
juvenile mussels are likely to be found) and then adjusting the criteria
to those pH conditions.  Another concern is whether or not the pH at the
appropriate depth is relatively constant.  These issues will likely vary
considerably depending on the characteristics of the sediment, and
perhaps seasonally.  The criteria can be adjusted to the pH in the
interstitial water to assess the potential for adverse effects at the
site.

Steven Canton	Because adult mussels siphon directly from the water
column, there is no need for a pH differential calculation.  However, to
be protective of the juvenile life stage which typically filters pore
water, there may be a need for differential pH measurements.  But how
does one account for all of the biotic and abiotic factors that affect
pH in the sediment?  Because site-specific sediment conditions greatly
influence the pH levels, adjusting the pH should only be an option for
site-specific calculations (or considered as “other data”) rather
than within the national water quality criteria.

Jerome Diamond	I don’t think criteria should consider sediment pH or
the difference between sediment and water pH in deriving water quality
criteria.  Rather, EPA should establish sediment ammonia criteria, using
toxicity test data for mussels and other benthic species (e.g.,
Hyalella, Hexagenia, etc.) exposed to sediments with ammonia.

George Dixon	My opinion here is partially covered in the response to
question 11. The pH of the sediment is only of consequence to the degree
that it influences the pH in the water column at the sediment-water
interface. I am not aware of any methodologies, other than direct
measurement, that would allow estimation of the interface pH if the only
data available were the sediment pH and the water-column pH.

Jerry Farris	In addition to the aforementioned comments regarding the
need for additional studies involving mussel responses before including
a consideration for this difference, site-specific variance procedures
inclusive of sediment characteristics for a “mussels present “ site
might be especially applicable in this instance.  Should these site
criteria account for such differences in testing of site waters with and
without sediment, they would be expected to be applicable for the
species (or suitable surrogate) and sediment of interest.  Use of a
site-specific criterion for copper that incorporated such broad test
considerations has been utilized to insure protection of mussel species
for almost two decades within the Clinch River, Virginia (Van Hassel,
2007).     

Bill Goodfellow	Same comment as for Question 11.

Jim Pletl	The freshwater ammonia criteria must appropriately address the
pH of the juvenile freshwater mussel environment to be reliable. 
Although the criteria must be protective they must also be
representative of aquatic life sensitivities and implemented in a way
that does not overstate those sensitivities.  Being protective is
straight-forward using conservative assumptions but criteria development
and implementation already have multiple layers of conservatism.  These
layers include the return frequency (once in 3 years) of the criteria,
the assumption that exposure is continuous instream for a time period
equivalent to the duration of the tests, an acute criterion averaging
period of one hour, the use of the 90th percentile of temperature and pH
to apply the ammonia criteria, the reasonable potential process of
EPA’s Technical Support Document for Water Quality-based Toxics
Control which relies on a very high percentile (95th or 99th) of the
distribution of the highest concentration measured for an effluent to
issue permit limits, the 303(d) listing process which is dependent on
the vast minority of data exceeding the criteria to conclude that a
water body is not in compliance with standards, etc.  The conservatism
built into criteria development and implementation demands that pH
instream be accurately assessed and applied to the criteria to reliably
determine compliance with water quality standards.  

EPA has not adopted data quality objectives to limit the amount of
uncertainty that can exist when applying water quality criteria
instream, therefore a consensus standard to determine if water based
data can be used to estimate sediment based responses is not available
and scientific defensibility cannot be assessed.  However, there is an
approach that will likely provide consensus agreement in scientific
defensibility: to directly address the pH of the instream sediment
within the context of a toxicity test rather than trying to compensate
for pH differences between water tests and the instream sediment.  This
approach will directly address concerns that juvenile test results are a
function of the presence/absence of sediment and avoid the uncertainty
of mechanisms attempting to extrapolate response from water only
exposures.  It is recognized that this recommendation would not allow
the use of current data to derive ammonia water quality criteria
inclusive of freshwater mussel sensitivities.  It would ensure that the
unique ecology of these mussels’ life stages is accurately captured
(within the constraints of conventional testing methods) and that when
the resulting criteria are implemented management actions will be
reliable. 

13.	Should exposure tests on juvenile mussels be conducted with or
without sediment in the test chamber?  

Alex Barron	Results of toxicity tests without sediment provide good
information regarding the toxicity of ammonia via exposure to ammonia
concentrations in water.  Acceptable control mortality in a test remains
a significant indicator that the test conditions and the general overall
health of the tested organisms are reasonable and the increasing
mortality observed in the test chambers with higher concentrations of
the tested substances is a direct result of the toxicity of the toxic
substance.  If the mortality in the control chambers remains low then,
this is good evidence that the tested organisms are not experiencing
unacceptable stress during the test.  In several water-only 28-day
tests, 100% to 90 % survival was observed in controls. This provides
evidence that water-only tests with juvenile mussels can provide
acceptable conditions for assessing the toxicity of a contaminant.  

Water-only exposure tests provide relatively comparable and reproducible
exposure conditions that are useful for assessing toxicity sensitivity
and to distinguish effects of pH temperature or hardness etc.  Toxicity
testing in sediment will not change what we know about the sensitivity
of the species to ammonia via a water-only exposure as used and
summarized in this criteria document. As shown in Tables 1, 2 and 4, the
variability of toxicity values within a species and within genera is all
within a factor of 10.  Water-only toxicity tests methods have shown
good intra and inter laboratory variability for toxicity tests with
ammonia, as well as copper and chlorine.  This lends confidence in these
data for use in criteria development.  Toxicity data from water-only
exposures are most relevant to the purposes that water quality criteria
are most often employed; assessments of natural waters by comparison to
the criteria and for use in developing regulatory limits for permitted
discharges of the pollutant.  

Concerns that tests conducted in sediment might produce different
toxicity results are speculation.  Some speculation is that water-only
exposure stresses the test organisms and may make the test organisms
more sensitive to the toxicant.  However the good survival in the
control chambers indicates that the tested mussels are not overly
stressed during the test.  Also, test conditions are designed to provide
optimum dissolved oxygen and temperature conditions.  Natural conditions
often are less than optimum with lowered oxygen levels and other
conditions.  The sediment may be a more natural environment for the
juvenile mussels, making them more hearty and resistant to the toxic
substance, or the sediment may act to absorb the toxic substance and
make it less bioavailable and less toxic.  On the other hand, if the
sediment may absorb the toxic substance and over time increase the
concentration in the interstitial pore water in relation to the
overlying water.   Most of these issues are site specific in nature and
beyond the scope of national recommended criteria. 

Natural sediments are highly variable and even a standardized
reconstituted “clean” sediment to be used in laboratory tests may
prove variable in some unknown way perhaps in response to some
unrecognized factor, however this is simply speculation.  Use of
sediment in toxicity tests will likely introduce added variability
without providing additional value for judging sensitivity and this will
introduce added complexity to interpreting the tests’ results for
criteria development.  

Several studies (Whiteman et. al., 1996, Hickey and Martin, 1999) have
reported that ammonia in interstitial pore water can often be higher
than concentrations in the water.     Overall, these concerns about
increased ammonia concentrations in interstitial pore water are not so
much related to toxicity sensitivity or issues that directly influence
the derivation of criteria.  These concerns and observations are more
related to an exposure assessment of potential risk to the sediment
dwelling organisms and the related physical factors that influence the
exposure conditions.

I believe that toxicity data from water-only tests that were conducted
under acceptable conditions (with acceptable control mortality) provide
the most useful data to judge the quantitative toxic effects of the
toxic substance.   These water-only data can then be used to assess
potential for toxicity in the water column.  Currently, there is
insufficient information available to indicate that the toxicity of
ammonia should be considered fundamentally different in interstitial
pore water as compared to the toxicity observed in water-only tests
under the same pH and temperature conditions.  An ammonia toxicity test
with water-only exposures provides reliable test results that can be
applied to exposures if the juvenile mussels are exposed to
concentrations of ammonia in water.    Water exposure tests should be
used in the derivation of water quality criteria and this information
should apply equally well to water column or interstitial pore water. 

If there is a special need to assess interstitial pore water at a site,
the criteria can be applied to an assessment of the interstitial pore
water.  Samples of interstitial water can be collected and analyzed for
ammonia concentrations and pH can be measured in the field.  If
monitoring of interstitial pore water is conducted, care should be taken
concerning the depth at which the samples are taken.  Juvenile mussels
are likely to be within a few centimeters of the surface and juveniles
and adult mussels are capable of some control over maintaining their
position in the sediment and should not be assumed to be regularly found
buried in deeper sediments where low dissolved oxygen, reduced food
availability and other stress factors may overshadow risk due to ammonia
toxicity.  Water samples for ammonia analysis and pH field measurements
should be taken at the surface interface or at a depth where mussels are
likely to be found. 

Steven Canton	If the test is designed to address water column concerns,
then the exposure test using juvenile mussels should be conducted
without sediment in the test chamber.  The use of sediment in water
quality toxicity tests introduces too many confounding factors in tests
designed to evaluate the needs to be protective of aquatic life in the
water column.  For many species of mussels, the juveniles primarily feed
on particles in the sediment and filter the pore water, rather than
directly siphon from the overlying water column.  Depending upon the
composition of the sediment (i.e., organic matter content) the sediment
layer can have profound influences on the pH and ammonia concentration
within the pore water of the sediment.  In addition, the metabolic
byproducts of juvenile organisms that live in sediments also influence
the level of ammonia observed within pore waters of sediment.  For these
reasons alone it is difficult to separate the effects solely due to
conditions within the water column versus the effects of the sediment
layer.

Furthermore, in static renewal tests that incorporate the use of
sediment, there is the potential for decreased ammonia concentrations,
as well as pH, in the pore water of sediment when compared to the water
column conditions (Allan and Maguire 1992); whereas in natural
conditions this relationship is often reversed, such that pH levels and
ammonia concentrations are greater in the sediments than the overlying
water column.   The sediment based toxicity studies of Whiteman et al.
(1996) tried to address these gradient issues between the sediment and
water column, by introducing nitrogen via sediment and maintaining
overlying water concentrations in a non-toxic range.  However, the
mobility of the test organisms allowed them to seek out the zones of
lower ammonia concentration at the interface rather than within the
sediment, which also confounded the test results.

This concludes my responses to the Charge Questions.  Given the time
constraints, I feel somewhat limited in what I could provide.  I hope
this review was helpful – and thank you for the opportunity.



Jerome Diamond	Tests using juvenile mussels should be conducted using
sediment.  Several researchers reported better survival and growth of
juvenile mussels with the addition of sediment (e.g., Gatenby et al.
1996, Yeager et al. 1994) and I would note that many of the mussel
culturing facilities that produce juveniles for transplants, use
sediment in their culturing systems.  The ASTM standard for glochidia
and juvenile mussel testing (E 24550-06) indicates this in several
places including sections 10.4.2.4, 10.5.2.1, 10.5.2.5, 10.6.3.3, and
10.6.3.13.  Laboratory studies by Yeager et al. (1994) observed that
young juveniles burrow immediately to < 1cm deep in the sediment, and
that they were not exposed to overlying water.  Others also have
reported rapid burrowing of new juveniles in sediment.  This information
suggests to me that some sediment in test containers may help ensure
that juvenile mussels are unstressed and respond appropriately to
contaminants.  This might be a good check to see whether ammonia
concentrations shown to be toxic in water-only exposures are also toxic
when the tests are conducted with some sediment (clean, sterile, etc.)
as well.

George Dixon	My comments here are limited to laboratory tests where
issues of pH variability between the sediment and the water column could
be eliminated (to avoid direct measurement as discussed above) by
experimental manipulation. The main reason for including sediment would
presumably be to reduce extraneous stress on the experimental organisms,
stress which could possibly interact with the toxicant to produce an
anomalous result. I have no opinion as to whether the presence or
absence of sediment would stress juvenile mussels in an experimental
setting. This will have to be determined experimentally. 

Jerry Farris	If the need is greater at this point with the reassessment
to consider data from tests that offer repeatability and precision of
results, than to ensure inclusion of test conditions reflecting the life
history and habitat selection for this life stage of the mussel, then
the test should exclude the sediment considerations.  There is a growing
body of evidence from tests without sediment that suggests comparable
juvenile shell growth can be measured in control treatments in static,
static-renewal, and flow-through conditions.  This is somewhat
contradictory to earlier studies that recognized the importance of
sediment in distinguishing effects during longer term exposures with
freshwater mussels and fingernail clams.  For this reason, the
information still seems lacking in determining how much of a confounding
effect is introduced by the effects of sediment characteristics (e.g.,
particle size, microbial mobilization and organic carbon content). More
specific information on the survival and physiological condition of
benthic organisms when tested in a water-only medium seems necessary for
testing beyond 96 hours.  Burton et al. (2002) describes a range of
toxicity assays using benthic invertebrates in sediment-free systems
such as interstitial water, elutriate phase, or spiked waters, with each
possibly differentiating toxicant uptake, pathway, and hazard.  It would
seem prudent to be consistent with current conventional approaches to
sediment evaluations for their consideration of appropriate phases to be
tested (extractable, elutriate, interstitial, whole sediment, and in
situ) relating to appropriate conditions for the species or substrate of
interest.  The complex interactions between mussel, water, and sediment
that influence their responses has been cited to reinforce those
considerations (Thorsen et al., 2007). 

Bill Goodfellow	I believe this question is very similar to the Hyalella
azteca question. If adequate test results that are representative to the
toxicity of ammonia can be performed in water only, that is how they
should be performed. If sediment is necessary for the test to be
performed adequately, then they should use sediment.  However, it should
be cautioned that when the variable of sediment is added to the toxicity
testing methodology, the complexity of the test increases considerably
as to the interpretation. In my opinion the agency should caution the
use of sediment in the test my lead to positive and negative bias of the
test results. The agency should try to stay with water column only
whenever possible.

Jim Pletl	It is clear from Newton and Bartsch (2007) that the potential
for juvenile freshwater mussels to be stressed in the absence of
sediment exists.  Wang et al. (2007) found that sediment did not appear
to impact shell growth and may have slightly reduced survival, but only
one test using sediment was conducted and this test only consisted of
two replicates per treatment.  As stated previously toxicity tests must
represent the stress experienced by a single stressor to be used in
developing water quality criteria.  Therefore for this particular use,
until more information is available, juvenile tests of any duration must
include sediment in the test vessels in order to be used in deriving
water quality criteria.

Additional Comments

Steven Canton	References:

Allan, G. L. and G. B. Maguire. 1992. Effect of pH and salinity on
survival, growth and osmoregulation in Penaeus monodon Fabricius.
Aquaculture 107:33-47.

Anderson, K.B., R.E. Sparks, and A.A. Paparo. 1978. Rapid Assessment of
Water Quality, Using the Fingernail Clam, Musculium transversum. WRC
Research Report No. 133. University of Illinois, Water Resources Center,
Urbana, IL.

Arid West Water Quality Research Project (AWWQRP).  2006.  Evaluation of
EPA Recalculation Procedure in Arid West Effluent-Dependent Waters:Final
Report. Prepared for the Arid West Water Quality Research Project by URS
Corporation, Chadwick Ecological Consultants, Inc., and Parametrix,
Inc., Pima County Wastewater Management Department, Tucson, AZ.

Arthur, J.W., C.W. West, K.N. Allen, and S.F. Hedtke. 1987. Seasonal
toxicity of ammonia to five fish and nine invertebrate species. Bull.
Environ. Contam. Toxicol. 38:324-331.

Augspurger, T., A.E. Keller, M.C. Black, W.C. Cope, and F.J. Dwyer.
2003. Water quality guidance for protection of freshwater mussels
(Unionidae) from ammonia exposure. Environmental Toxicology and
Chemistry 22:2569-2575.

Bartsch, M.R. et al. 2003. Effects of pore water ammonia on in situ
survival and growth of juvenile mussels (Lampsilis cardium) in the St.
CroixRiverway. Environ Toxicol Chem 22:2561-2568

Calamari, D., R. Marchetti, and G. Vailati. 1977. Effects of prolonged
treatments with ammonia on stages of development of Salmo gairdneri.
(English Translation) Nuovi Annali d’Igiene e Microbiologia.
28:333-345.

Colt, J.E., and G. Tchobanoglous. 1978. Chronic exposure of channel
catfish, Ictalurus punctatus, to ammonia: effects on growth and
survival. Aquaculture 15:353-372.

Dankers N, K. Koelemaij. 1989. Variations in the mussel population of
the Dutch Wadden Sea in relation to monitoring of other ecological
parameters.  HYPERLINK
"http://www.springerlink.com/content/103796/?p=265c1a04a27c42bdaafba7b3b
6b9a9b9&pi=0" Helgoland Marine Research   HYPERLINK
"http://www.springerlink.com/content/t07111844315/?p=265c1a04a27c42bdaaf
ba7b3b6b9a9b9&pi=0" 43:  529-535

Khatami, S. H., D. Pascoe, M. A. Learner.  1998.  The acute toxicity of
phenol and unionized ammonia, separately and together, to the
ephemeropteran Baetis rhodani (Pictet). Environmental Pollution
99:379-387.

Mount, D.I. 1982. Ammonia toxicity tests with Ceriodaphnia acanthine and
Simocephalus vetulus. Memorandum to Dr. R.C. Russo, ESEPA, August 6,
1982.

Reinbold, K.A., and S.M. Pescitelli. 1982. Acute Toxicity of Ammonia to
the White Sucker. Final Report to EPA. Center for Aquatic Ecology,
Illinois Natural History Survey, Champaign, Illinois.

Rubin, A.J., and M.A. Elmaraghy. 1977. Studies on the toxicity of
ammonia, nitrate and their mixtures to guppy fry. Water Research
11:927-935.

Russo, R.C., D.J. Randall, and R.V. Thurston. 1988. Ammonia toxicity and
metabolism in fishes. In R.C. Ryans, ed. Protection of River Basins,
Lakes, and Estuaries. pp. 159-173. American Fisheries Society, Bethesda,
Maryland.

Sangli, A. B. and V. V. Kanabur. 2001. Toxicity of ammonia to a
freshwater fish, Gambusia affinis and its effect on oxygen consumption.
Geobios 28:56-58.

Sparks, R.E. and M.J. Sandusky. 1981.  Identification of Factors
Responsible for Decreased Production of Fish Food Organisms in the
Illinois and Mississippi Rivers.  Final Report for Project No. 3-291-R,
Illinois Natural History Survey, River Research Laboratory, Havana,
Illinois.

Stephan, C.E., D.I. Mount, D.J. Hansen, J.H. Gentile, G.A. Chapman, and
W.A. Brungs.  1985.  Guidelines for Deriving Numerical National Water
Quality Criteria for the Protection ofAquatic Organisms and Their Uses. 
PB-85-227049.  U.S. Environmental Protection Agency, Office of Research
and Development, Duluth, Minnesota.

Thurston, R.V., R.J. Luedtke, and R.C. Russo. 1984. Toxicity of Ammonia
to Freshwater Insects of Three Families. Technical Report No. 84-2,
Fisheries Bioassay Laboratory, Montana State University, Bozeman, MT. 26
pp.

Thurston, R.V., R.C. Russo, and G.A. Vinogradov. 1981. Ammonia Toxicity
to Fishes. Effect of pH on the Toxicity of the Un-ionized Ammonia
Species. Environ. Sci. Technol. 15:837-840.

Tomasso, J.R., C.A. Goudie, B.A. Simco, and K.B. Davis. 1980. Effects of
environmental pH and calcium on ammonia toxicity in channel catfish.
Transactions of the American Fisheries Society 109:229-234.

U.S. Environmental Protection Agency (USEPA). 1999 Update of Ambient
Water Quality Criterion for Ammonia. Office of Water, Washington, DC.
EPA-822-R-99-014.

U.S. Environmental Protection Agency (USEPA).  2001.  2001 Update of
Ambient Water Quality Criteria for Cadmium.  EPA-822-R-01-001.  Office
of Water, Washington, DC.

Yeager, M.M., D.S. Cherry and R.J. Neves. 1994. Feeding and burrowing
behavior of juvenile rainbow mussels, Villosa iris (Bivalvia:
Unionidae). J.N. Am. Benthol. Soc. 13(2):217-222.

Wallen, I.E., W.C. Greer and R. Lasater. 1957. Toxicity to Gambusia
affinis of certain pure chemicals in turbid waters. Sewage Ind. Wastes.
29: 695-711.

Whiteman, F.W., G.T. Ankely, M.D. Kahl, D.M. Rau, and M.D. Balcer. 1996.
Evaluation of interstitial porewater as a route of exposure for ammonia
in sediment tests with benthic macroinvertebrates. Environmental
Toxicology and Chemistry. 15(5):794 - 801.

Jerome Diamond	Other Comments Regarding Appendex A

Under “USEPA’s Aquatic Life Criteria Coordinating Committee”
comments, point #2 makes it appear that EPA is willing to alter a
national criterion due to a commercially valuable species (e.g., trout)
but not an ecologically critical species?  Point #3:  I agree there is
no evidence that glochidia in conglutinates are exposed to water-borne
pollutants but there’s no evidence that they are not exposed either. 
Also, uptake of pollutants by planktonic invertebrates lacking gills
(e.g., Daphnia) occurs via absorption through the carapace, not just
through food.  Point #8:  “For glochidia that attach to hosts,” –
all glochidia are meant to attach to hosts.  What I think you mean here
is “In terms of the duration of free-living glochidia, there is a
major difference….”

Other comments regarding acute data in the Reassessment Document

Daphnia GMAV is Incorrect

I believe the document reports an incorrect GMAV for Daphnia (p. 9) as
well as test data in Table 1 (p. 49-50).  It appears that the unionized
ammonia values reported by Russo et al. (1985) and reported as such in
EPA’s 1999 document, were incorrectly transcribed as total ammonia
values (see p. 109, Appendix 4 of 1999 document as well as p. 38, Table
4 of the 1999 document).  Even with the temperature correction for
invertebrates in the Reassessment Document, Daphnia couldn’t possibly
go from Rank 19 in acute sensitivity to 8th in the draft document.  This
error has other ramifications as well:  assuming the Daphnia value is
incorrect in the draft, when mussels were absent, the four most
sensitive genera are Corbicula, Potamopyrous, Pleurocera, and Prosopium
(mountain whitefish).  If Corbicula is removed along with the unionids
as EPA proposed (and which I disagree with), then Deltistes (Lost River
sucker) becomes the 4th most sensitive genera.  Either way,
invertebrates no long occupy the four most sensitive genera.  Although I
did not do a comprehensive check, it appears that the other acute values
in the draft document are consistent with the 1999 document.

Temperature Equation for Invertebrate and Acute Data should be
re-examined

The temperature-ammonia sensitivity relationship applied by EPA to
invertebrate acute data in the draft document is based on the slope
provided in the 1999 document, based on Arthur et al. (1987), but not
used for the acute criterion in that document.  As the 1999 document
makes clear, the temperature relationships from Arthur et al. (1987)
were based on seasonal tests of experimental stream water, including
condition, life stage, etc. of test organisms collected during each
season.  As Arthur et al. (1987) notes, several water quality
characteristics other than temperature varied with season and, even more
importantly perhaps, organism condition was unlikely to be similar
across seasons and at the least was unknown.  EPA provided what I
believe was a poor rationale for using Arthur’s et al. data for
chronic invertebrate data.  The fact that Arthur did not observe a clear
temperature relationship with fish (which are also cold blooded) would
weaken the case that temperature was what was driving the apparent
ammonia sensitivity in invertebrates in his study (and which Arthur
himself acknowledges in his paper).

The temperature-toxicity relationship derived by EPA for ammonia in the
1999 document had a fairly steep slope; e.g., a difference between 20
and 25( C resulted in 2-3 fold decrease in the LC50.  There are a few
invertebrate species in the draft document for which it appears possible
to reanalyze a temperature toxicity relationship, and to determine if
one actually exists.

I have not had time to do a formal analysis of the invertebrate acute
data, however, my initial reaction is that temperature may be somewhat
weakly related to acute ammonia toxicity in invertebrates.  For example,
for the snail species Potamopyrgus antipodarum, Hickey and Vickers
(1994) reported an LC50 of 8.71 mg N/L at pH = 8.2 and 15( C and an LC50
of 4.08 mg N/L at pH = 8.2 and 25( C (a factor of 2 over a 10( C
difference).  The LC50 they reported at the same pH and 20( C was nearly
identical to the LC50 at 25( C (4.73 mg N/L).  LC50’s for the
fatmucket mussel at pH = 8.1 and 20( C ranged from 5.2 mg N/L to 11 mg
N/L (Wang et al. 2008) a factor of about 2, which could be within the
range specified by a 5( C difference in EPA’s temperature equation. 
Even larger differences in LC50’s are commonly reported for other
invertebrates based on tests at the same pH and temperature (e.g.,
Villosa iris LC50 = 3.0 – 8.9 mg N/L at pH = 8.3 and 20( C).  These
data suggest that a temperature adjustment factor may be premature for
acute invertebrate data.  Acute data for Ceriodaphnia dubia may perhaps
provide the best information for invertebrates thus far.  At pH = 8.2
and temperature of 7( C, the LC50 = 16.65 mg N/L (Nimmo et al. 1989). 
At pH = 8.16 and 22( C, the LC50 = 30.08 mg N/L (Black 2001), however,
at pH = 8.0 and 25( C, the LC50 = 14.52 mg N/L (Scheller 1997), similar
to the LC50 reported at 7( C.  EPA should reanalyze current acute data,
similar to the way they did for the 1999 document, to determine the
relationship between temperature and acute ammonia toxicity to
invertebrates.

LITERATURE CITED

Arthur, J.W., C.W.West, K.N. Allen and S.F. Hedtke. 1987. Seasonal
toxicity of ammonia to five fish and nine invertebrates species. Bull.
Environ. Contam. Toxicol. 38(2): 324-331.

Black, M. 2001.Water quality standards for North Carolina's endangered
mussels. Department of Environmental Health Science, Athens, GA.

Burton, G.A.  1992.  Sediment collection and processing factors
affecting realism.  In:  G.A. Burton, Ed., Sediment Toxicity Testing. 
Lewis, Boca Raton, FL, pp. 37-64.

Catherine M. Gatenby, Richard J. Neves and Bruce C. Parker 1996. 
Influence of Sediment and Algal Food on Cultured Juvenile Freshwater
Mussels.  Journal of the North American Benthological Society, Vol. 15: 
597-609

Dodd, BJ, MC Barnhart, CL Rogers-Lowery, TB Fobian, RV Dimock, Jr, 2006.
Persistence of host response against glochidia larvae in Micropterus
salmoides. Fish & Shellfish Immunology 21:473-484

Hickey, C.W. and M.L. Vickers. 1994. Toxicity of ammonia to nine native
New Zealand freshwater invertebrate species. Arch. Environ. Contam.
Toxicol. 26(3): 292-298.

Kernaghan, N. J., Gross, T. S., Bishop, C.D., Wang, N., Roberts, A., and
Ingersoll, C. G.. 2005. “Laboratory Toxicity Testing with Freshwater
Unionid Mussels,” Freshwater Bivalve Ecotoxicology, van Hassel, J. H.,
and Farris, J. L., eds., SETAC Press, Pensacola, FL.

M. M. Yeager, D. S. Cherry and R. J. Neves. 1994. Feeding and Burrowing
Behaviors of Juvenile Rainbow Mussels, Villosa iris (Bivalvia:Unionidae)
Journal of the North American Benthological Society, Vol. 13: 217-222

Neves, R. J., and Widlak, J. C. 1988. Occurrence of Glochidia in Stream
Drift on Fishes of the Upper North Fork Holston River, Virginia.  Am
Midland Naturalist, 119, 1988, pp. 111-120.

Nimmo, D.W.R., D. Link, L.P. Parrish, G.J. Rodriguez, W. Wuerthele and
P.H. Davies. 1989. Comparison of on-site and laboratory toxicity tests:
Derivation of site-specific criteria for un-ionized ammonia in a
Colorado transitional stream. Environ. Toxicol. Chem. 8(12): 1177-1189.

Rogers-Lowery, CL and RV Dimock, Jr.  2006. Encapsulation of attached
ectoparasitic larvae of freshwater mussels by epithelial tissue on fins
of naive and resistant host fish. Biol. Bull. 210:51-63 

Salmon, A. and R.H. Green.  1983.  Environmental determinants of unionid
clam distribution in the Middle Thames River, Ontario.  Canadian Journal
of Zoology 61: 832-838.

Sardia, N. and G. Burton.  1995.  Ammonia variation in sediments: 
Spatial, temporal and method-related effects.  Environ. Toxicol. Chem.
14(9):1499-1506.

Scheller, J.L. 1997. The effect of dieoffs of Asian clams (Corbicula
fluminea) on native freshwater mussels (unionidae). Virginia Polytechnic
Institute and State University, Blacksburg, VA.

Schwal, A. and M. Pusch  2007. Horizontal and vertical movements of
unionid mussels in a lowland river Journal of the North American
Benthological Society 26(2):261–272.

Stemmer, B.L., G.A. Burton, Jr. and G. Sasson-Brickson.  1990.  Effect
of sediment spatial variance and collection method on cladoceran
toxicity and indigenous microbial activity determinations.  Environ.
Toxicol. Chem. 9:1035-1044.

Strayer, D. L., Downing, J. A., Haag,W. R., King, T. L., Layzer, J. B.,
Newton, T. J., and Nichols, S. J. 2004. Changing Perspectives on Pearly
Mussels, North America’s Most Imperiled Animals. BioScience, 54:2004,
pp. 429-439.

U.S. Environmental Protection Agency.  1994.  Interim Guidance on
Determination and Use of Water Effect Ratios for Metals, Appendix B: The
Recalculation Procedure.  EPA-823-B-94-001.  Office of Water,
Washington, DC.

USEPA. 2000.  Ambient Aquatic Life Water Quality Criteria for Dissolved
Oxygen (Saltwater):  Cape Cod to Cape Hatteras.  EPA-822-R-00-012 Office
of Water, Washington, D.C. November 2000 

USEPA. 2002.  Clinch and Powell Valley Watershed Ecological Risk
Assessment. 2002.  EPA-600-R-01-050.  NCEA, ORD, Washington, D.C.

Wang, N., C.G. Ingersoll , D.K. Hardesty, C.D. Ivey, J.L. Kunz, T.W.
May, F.J. Dwyer, A.D. Roberts, T. Augspurger, C.M. Kane, R.J. Neves and
M.C. Barnhart. 2007. Contaminant sensitivity of freshwater mussels:
Acute toxicity of copper, ammonia, and chlorine to glochidia and
juveniles of freshwater mussels (Unionidae). Environ. Toxicol. Chem.
26(10):2036-2047.

Wang, N., J.T. Augspurger, M.C. Barnhart, J.R. Bidwell, WG. Cope, F.J.
Dwyer, S. Geis, I.E. Greer, C.G. Ingersoll, C.M., Kane, T.W. May, R.J.
Neves, T.J. Newton, A.D. Roberts and D.W. Whites. 2007. Contaminant
sensitivity of freshwater mussels: Intra- and inter-laboratory
variability in acute toxicity tests with glochidia and juveniles of
freshwater mussels (Unionidae). Environ. Toxicol. Chem.
26(10):2029-2035.

Watters, G. A., 2005. A Brief Look at Freshwater Mussel (Unionacea)
Biology. Freshwater Bivalve Ecotoxicology, van Hassel, J. H., and
Farris, J. L., eds., SETAC Press, Pensacola, FL.

Watters, G. T., S. H. O’dee, And S. Chordas. 2001. Patterns of
vertical migration in freshwater mussels (Bivalvia: Unionidae). Journal
of Freshwater Ecology 16:541–549.

Way, C.M., A.C. Miller and B.S. Payne.  1990.  The influence of physical
factors on the distribution and abundance of freshwater mussels
(Bivalvia:Unionidae) in the lower Tennessee River.  The Nautilus 103(3):
96-98.

George Dixon	I have reviewed the document listed above as requested in
the Charge to Reviewers dated 24 July 2009. In addition I have gone
through the background document “1999 Update of Ambient Water Quality
Criteria for Ammonia”. Let me state at the outset that I find the
Draft Reassessment to be a well conceived, well developed and carefully
written document. The acute and chronic ammonia criteria for freshwaters
with mussels present, and separate criteria for waters with mussels
absent, are well founded and scientifically defensible. I have presented
my review comments as responses to the thirteen questions posed in the
Charge to Reviewers Those questions are repeated here for clarity. 

Jerry Farris	References:

American Society for Testing and Materials. 2006. Standard guide for
conducting laboratory toxicity tests with freshwater mussels. E2455-06.
In Annual Book of ASTM Standards, Vol 11.06. Philadelphia, PA.

Augspurger, T. 2007.  Advances and opportunities in assessing
contaminant sensitivity of freshwater mussel (Unionidae) early life
stages.  Environ Toxicol Chem 26:2025-2028.

Bartsch MR, Newton TJ, Allran JW, O’Donnell JA, Richardson WB. 2003.
Effects of pore-water ammonia on in situ survival and growth of juvenile
mussels (Lampsilis cardium) in the St. Croix

riverway, Wisconsin, USA. Environ Toxicol Chem 22:2561–2568.

Burton, GA, PM Chapman, and EP Smith.  2002.  Weight of evidence
approaches for assessing ecosystem impairment.  Human and Ecological
Risk Assessment: An international Journal 8:1657-1673.

 

Cherry, DS, JL Scheller, N.L. Cooper and J.R. Bidwell. 2005. The
potential for Asian clam (Corbicula fluminea) dieoffs to impact native
freshwater mussels (Unionidae) I: Water column ammonia levels and
ammonia toxicity. Journal of the North American Benthological Society,
24:369-380.

Cooper, NL. JR Bidwell and DS Cherry. 2005. The potential for Asian clam
(Corbicula

fluminea) die-offs to impact native freshwater mussels (Unionidae) II:
Pore-water ammonia. Journal of the North American Benthological Society,
24:381-394. 

Mummert, AK, Neves RJ, Newcomb TJ, Cherry DS. 2003. Sensitivity of
juvenile freshwater mussels (Lampsilis fasciola, Villosa iris) to total
and un-ionized ammonia. Environ Toxicol Chem 22:2545–2553.

Newton TJ, Allran JW, O’Donnell JA, Bartsch MR, Richardson WB. 2003.
Effects of ammonia on juvenile unionid mussels (Lampsilis cardium) in
laboratory sediment toxicity tests. Environ Toxicol Chem 22:2554–2560.

Salazar, MH and SM Salazar.  2007.  Linking bioaccumulation and
biological effects to chemicals in water and sediment:  A conceptual
framework for freshwater bivalve ecotoxicology.  In Farris JL, Van
Hassel JH, eds, Freshwater Bivalve Ecotoxicology.  SETAC, Pensacola, FL,
USA, pp 215-250.

Scheller JL. 1997. The effect of die-offs of Asian clams (Corbicula
fluminea) on native freshwater mussels (Unionidae). Masters thesis.
Virginia Polytechnic Institute and State University, Blacksburg, VA,
USA.

Stephan CE, Mount DI, Hansen DJ, Gentile JH, Chapman GA, Brungs WA.
1985. Guidelines for deriving numerical national water quality criteria
for the protection of aquatic organisms and their uses. EPA 822/A85/100.
U.S. Environmental Protection Agency, Office of Research and
Development, Washington, DC.

Thorsen, WA, WG Cope, and D Shea.  2007.  Toxicokinetics of
environmental contaminants in freshwater bivalves.  In Farris JL, Van
Hassel JH, eds, Freshwater Bivalve Ecotoxicology.  SETAC, Pensacola, FL,
USA, pp 169-207. 

Van Hassel, JH.  2007.  Case study: Discrimination of factors affecting
Unionid mussel distribution in the Clinch River, Virginia, USA.  In
Farris JL, Van Hassel J, eds, Freshwater Bivalve Ecotoxicology.  SETAC,
Pensacola, FL, USA, pp 311-331.

Wang N, Ingersoll CG, Hardesty DK, Ivey CD, Kunz JL, Dwyer FJ, Roberts
AD, Augspurger T, Kane CM, Neves RJ, Barnhart MC. 2007. Acute toxicity
of copper, ammonia, and chlorine to glochidia and juveniles of
freshwater mussels (Unionidae). Environ Toxicol 26:2048-2056.

Wang N, Ingersoll CG, Greer IE, Hardesty DK, Ivey CD, Kunz JL, Dwyer FJ,
Roberts AD, Augspurger T, Kane CM, Neves RJ, Barnhart MC. 2007. Chronic
toxicity of copper and ammonia to juvenile freshwater mussels
(Unionidae). Environ Toxicol Chem 26:2048-205.

Bill Goodfellow	Re:  Review of EPA’s Draft Reassessment of the 1999
Ambient Water Quality Criteria for Ammonia-Freshwater.

Dear Ms. Waite:

I have found the criteria document to be well organized and very well
thought out.  I also feel that the scientific justification laid out in
the document to be excellent.

I have several questions or comments presented below as well as imbedded
in the Charge Questions section (my comments are presented in Bold).

I do question why only standards consistent with ASTM Standards for
testing (page 14) are the only relevant test procedures that were used
in the document. I believe that the EPA’s Whole Effluent Testing (WET)
toxicity testing guidelines for testing are also relevant. Did this
exclude any additional data? From my review of the references, I do not
believe so.

I also question when the Nile tilapia was place in the other chronic
toxicity data (and I agree this was appropriate), why is the Asiatic
clam included in the data used for development of the criteria (and in
fact it is the third most sensitive species for the acute criterion).
The Asiatic clam is also a nonindigenous species (or exotic). In my
opinion I don’t believe that the Asiatic clam should not be used in
the development of the criteria regardless that this exotic is found
throughout United States. I believe elimination of this species from the
dataset will still generate a protective acute criterion.

The Table A in text of the GMAV for the Mountain Whitefish does not
match Table 1. The Text table reports 12.11 and Table 1 reports 12.09. 
I am expecting that the Table 1 is correct and is what was used for the
FAV. Please verify.

I thank you for the opportunity to allow me to be a peer reviewer for
this important freshwater water quality criteria document. If you have
any questions or require additional information, please let me know.

Sincerely,

William L. Goodfellow, Jr.

Vice President and Senior Scientist

Clarification question from EPA for the following comment made by Dr.
Bill Goodfellow:

"I also question when the Nile tilapia was place in the other chronic

toxicity data (and I agree this was appropriate), why is the Asiatic

clam included in the data used for development of the criteria (and in

fact it is the third most sensitive species for the acute criterion).

The Asiatic clam is also a nonindigenous species (or exotic). In my

opinion I don't believe that the Asiatic clam should not be used in the

development of the criteria regardless that this exotic is found

throughout United States. I believe elimination of this species from the

dataset will still generate a protective acute criterion."

EPA Question: “…it sounds like he's questioning why we included the
clam

data but then the text I highlighted in red sounds like he's saying that

we should include it - maybe there's a typo and the double negative

sentence is wrong?  If you could get clarification from him for me on

this I would appreciate it.

Clarification from Dr. Bill Goodfellow:

In my opinion the Asiatic clam (an exotic) should not be included in the
database used for development of the criteria.  Rather it should be
placed in the Table with other data.  If the Agency decides to include
the Asiatic Clam in the criteria development, I believe considerable
justification will need to be provided in the written justification of
the criteria document.

Hopefully this makes it clearer. Let me know if you have any other
questions or require additional information.



Jim Pletl	LITERATURE CITED

R.B Bringolf, W.G. Cope, S. Mosher, M.C. Barnhart, D. Shea. 2007a. 
Acute and Chronic Toxicity of Glyphosate Compounds to Glochidia and
Juveniles of Lampsilis siliquoidea (Unionidae).  Env. Tox. Chem.26(10),
p.2094

R.B Bringolf, W.G. Cope, M.C. Barnhart, S. Mosher, M.C. Barnhart, P.
Lazaro, D. Shea. 2007b.  Acute and Chronic Toxicity of Pesticide
Formulations (Atrazine, Choropyrifos and Permethrin) to Glochidia and
Juveniles of Lampsilis siliquoidea (Unionidae).  Env. Tox. Chem.26(10),
p.2101

T.J Newton and M.R. Bartsch. 2007.  Lethal and Sublethal Effects of
Ammonia to Juvenile Lampsilis Mussels (Unionidae) in Sediment and Water
Only Exposures.  Env. Tox. Chem. (26(10), p.2057

N. Wang, C.G. Ingersoll, I.E. Greer, D.K. Hardesty, C.D. Ivey, J.L.
Kunz, W.G. Brumbaugh, F.J. Dwyer, A.D. Roberts, T. Augspurger, C.M.
Kane, R.J. Neves, M.C. Barnhart.  2007.  Chronic Toxicity of Copper and
Ammonia to Juvenile Freshwater Mussels (Unionidae).  Env. Tox. Chem.
26(10), p. 2048





SECTION C

Individual Peer Reviewer Comments



PEER REVIEW COMMENTS FROM

Alex Barron

Virginia Department of Environmental Quality

Office of Water Quality Criteria and Standards

Richmond, VA 

EPA’s Draft Reassessment of the 1999 Ambient Water Quality

Criteria for Ammonia – Freshwater

July 14, 2009

Review Comments by 

Alex Barron

Virginia Department of Environmental Quality

August 18, 2009 

CHARGE QUESTIONS

Acute criteria in fresh waters:

Question:

Are the toxicity tests used to derive the criteria scientifically
defensible for such use?  Are you aware of other relevant data that were
not used?

 Review Comment:

	Yes, all the toxicity tests that influence the derivation of the new
criteria for ammonia appear to be scientifically defensible. All
relevant data that I am aware of has been considered in this
reassessment document. 

Question:

What are the technical considerations that EPA should evaluate when
mussels are present and mussels are absent with respect to the
recommended acute criteria?

Review Comment:

	Table B compares Final Acute Value (FAV) and Criterion Maximum
Concentration (CMC) values calculated with the entire dataset including
data for bivalve mollusks and also with the dataset with the data for
bivalve mollusks removed.  The Asiatic clam (Corbicula fluminea) is in a
different Family (Cobiculidae) and Order (Veneroida), and is not a
Unionid mussel.  Corbicula has shown very similar sensitivity to ammonia
compared to Unionid mussels however.  Table B really represents a
non-bivalve mollusk dataset, which would be appropriate where no bivalve
mollusks of any kind are present.  The FAV and CMC calculated without
the data for bivalve mollusks is less than a factor of two different
than the FAV using the entire dataset, which is 62.75 % higher.   The
1999 ammonia criteria also provided separate recommended acute criteria
based on observed sensitivity of salmonid species.  The 1999 criteria
document also presented separate acute criteria that are approximately
50 % different between the salmonids present or absent acute criteria.
The recommendations in the 2009 reassessment document providing separate
criteria , calculated with and without a group of sensitive species is
in keeping with precedents set in the earlier ammonia criteria for
calculating separate criteria when one group of species appears to be
especially sensitive to ammonia.    This can be considered appropriate
for ammonia criteria because of the relatively large database and the
fact that ammonia toxicity is usually caused by mg/L concentrations as
compared to ug/L concentrations as seen with many other toxic
pollutants. 	 

	The implementation of criteria based on the dataset without freshwater
mussels will require some way of determining whether or not a water body
is not likely to contain freshwater mussels, before such criteria can be
considered appropriate.  The presence or absence of freshwater mussels
in a water body may be difficult to ascertain without a survey that
specifically targets this issue.  Most monitoring techniques for
collecting benthic macroinvertibrates such as the widely used EPA Rapid
Bioassessment Protocol (RBP) are not designed or intended to regularly
collect mussels that are often tightly imbedded in the sediment.  Thus a
typical RBP survey that did not collect freshwater mussels cannot be
accepted as proof of absence of mussels in that waterbody.  I would
recommend a default of assuming the presence of freshwater mussels
unless several surveys that specifically targeted the identification of
mussels have been conducted on that water body.  A brief review of
biological survey data in Virginia showed that the presence of
freshwater mussels was detected in 83 % of water bodies that had been
surveyed with methods capable of finding mussels.  However, in some
waterbodies the presence of mussels was confirmed only after several
surveys. Because the distribution of freshwater mussels is often
dependent on fish for dispersal, it might be suspected that mussels may
not be present in some headwater or intermittent steams where it can be
demonstrated that fish are not present due to physical barriers
preventing fish migration.  

Question:

Is it scientifically defensible to exclude the glochidia data at this
time due to the uncertainty of appropriate test duration time for this
life stage?  Do you believe there is an alternative approach to the use
of this data that would be more scientifically sound?

Review Comment:

	The position statement in Appendix B of the reassessment document
acknowledges several concerns about the use of toxicity data for
glochidia and recommends that these data should not be used for criteria
development at the present time.   This position seems reasonable given
the uncertainties regarding glochidia testing as outlined in Appendix B.
Glochidia, with such short-lived life stages and with variable lengths
of natural viability in the water column, ranging from a few hours to
several days, it is difficult to know how to use some of these data for
developing criteria.   For some species where the natural viability of
glochidia, or where the actual exposure to the water column is short,
the result of 48 hour tests may not provide a realistic exposure
scenario, while a shorter exposure may be more important..   

	There are some additional short-term (6 and 24-hour) exposure data
available that may provide some additional information on this issue.
Table 2 of the Reassessment Document provides acute values based on
48-hour exposures for tests conducted with glochidia in several
publications.   For some of these same tests, additional information on
6 hour and 24-hour exposures are also contained in Wang et al 2007b, in
their table 3.  These 6 and 24-hour data indicate that glochidia are
less sensitive to ammonia during 6 and 24-hour exposures than during a
48 hour exposure. For the species where glochidia may only be viable or
exposed to the water column for a few hours, the results of 6 hour and
24 hours tests may provide useful information and these data could be
normalized to the proper standard conditions and included in Table 2, to
provide additional information for comparison.  As reported in Wang et
al 2007b; the data suggest that at 6 to 24 hour test durations, the
sensitivity of glochidia for most species of freshwater mussels is less
than or equal to the sensitivity of the juveniles  of the same species
at 96 hours.   This provides some support that the revised ammonia acute
criteria (based on 96 hour data for juveniles) may also provide
protection to glochidia at exposure durations of 24 hours or less.   The
fact the acute criterion is applied as a one-hour average should also
provide some additional level of protection.

Question:

Regarding the proposed approach to glochidia data in the 2009 draft
position statement as it relates to ecological relevance and
practicality - Is the approach a scientifically defensible principle for
structuring the population exposure duration problem and designing
further research to quantify such a duration?

Review Comment:

	I believe that the approach outlined in appendix B is defensible. 
Given the significant uncertainties concerning the appropriate exposure
durations for glochidia as discussed in the appendix B, attempting to
use toxicity data currently available for glochidia to derive defensible
criteria at the present time appears to be untenable.  This would be
especially important if the final criteria were significantly influenced
by glochidia data with the recognized uncertainties about practical
ecological importance.  Water quality criteria should be based on the
best information available in order to gain acceptance and avoid
continued uncertainty.  Additional information as discussed in appendix
B would be needed to provide sufficient confidence in utilizing
glochidia data to derive water quality criteria.     



Hyalella azteca position statement and proposed rationale (see Appendix
B):

The EPA workgroup developed a position statement and proposed supporting
rationale describing the concerns over using Hyalella azteca toxicity
test data in criteria development due to the uncertain health of the
test organisms in different test water composition.  The rationale
defines the specific concerns and uncertainties supporting the
recommended exclusion of the Hyalella data from use in criteria
derivation, at this time; the position statement is based on the
workgroup’s review of a number of toxicity tests on Hyalella,
referenced in the rationale.  

Question:

5.	Are the position statement and supporting rationale regarding use of
toxicity data for Hyalella azteca in criteria development reasonable and
scientifically sound recommendations?  

Review Comment:

Yes.  The EPA Guidelines for Diving Numerical National Water Quality
Criteria for the Protection of Aquatic Organisms and Their Uses, in
section IV. 5. H specifies that in order to be useful for calculating a
Final Acute Value, the must be relatively good agreement of the data
within a species.  Generally “if the acute values differ by more than
a factor of 10, some or all of the values probably should not be used in
the calculation” of a Species Mean Acute Value (SMAV).    This applies
to the available data set for Hyalella azteca, where the normalized
acute values listed in Table 2 range from 1.58 mg/L to 83.9 mg/L,
differing by as much as a factor of 53.  This wide range of acute values
for H. azteca required more careful review before determining what if
any of the data for this species is acceptable for criteria development.
  EPA has conducted an additional review of these data and has
determined that there are significant uncertainties with data for this
species, enough to preclude their use in deriving criteria.  The
position statement provides a summary of the information known
concerning possible factors that could cause the acute toxicity values
for H. azteca to be affected by some water quality characteristics not
previously recognized or considered and not fully understood at this
time.  Based on the current information available, it appears reasonable
that until these factors are better understood, the data for H. azteca
should not be used in the calculation of a SMAV or a FAV.



  SEQ CHAPTER \h \r 1 Chronic Freshwater Criterion:

Question:

6.  Are the toxicity tests and other studies used to derive the
criterion scientifically defensible for such use?  Are you aware of
other relevant data that were not used?

Review Comment:

Yes.  The toxicity studies referenced all appear to be good quality
studies and provide useful information on the chronic toxicity of
ammonia to aquatic organisms.  The use of EC20 values calculated by
regression analysis is excellent and provides a consistent estimation of
low-level effects under chronic conditions. 

All toxicity data relevant to ammonia toxicity that I am aware of are
included in the references and were considered in the reassessment
document.  

Question:

7.	Is the freshwater chronic criterion scientifically defensible with
mussels present and mussels absent?  

Review Comment:

Yes.  Similar distinctions for the presence or absence of certain
sensitive organisms showing different sensitivities to ammonia have been
used before.  The 1999 ammonia criteria recommended different chronic
criteria based on the presence or absence of early life stages of fish
in the water body, setting a precedent of allowing different criteria
for different situations.  These different criteria were based on a
similar magnitude of difference in sensitivity between the early life
stages and the older life stages of fish (in the 1999 reassessment
criteria) as is seen between the sensitivities of freshwater mussels and
other species (in the 2009 reassessment).  This policy of providing
these separate criteria based on the presence absence of freshwater
mussels will provide some flexibility in implementing the criteria in
some instances and is reasonable.  

Given the wide spread nature of ammonia in natural environments and the
difficulties and expense imposed on municipalities in treating for
ammonia in sewage treatment facilities, a significant lowering of the
ammonia criteria is likely to have important consequences for
municipalities.  Providing these two criteria may help alleviate
concerns that any additional restrictions on ammonia discharges are
targeted to providing the needed level of protection for the waterbody,
based on whether or not the more sensitive mussels are present.  
Providing two EPA recommended criteria that already take these issues
into account could help reduce the potential for requests for
consideration of developing site-specific criteria based on the
recalculation procedure, which can require significant resources to
pursue.   

Table C shows the four most sensitive genera and the calculation of the
CCC for the two different criteria (with and without Unionid mussel
data).  The data under “excluding freshwater mussel data (family
Uniondale)” includes the GMCV for the fingernail clam, Musculium
transversum as the third most sensitive genera and this GMCV is used in
the calculation of the CCC for “non-Unionid containing waters”. 
While it is true that M. transversum species is not in the family
Unionidae, it is a bivalve mollusk.  

In the recommendations for  developing separate acute criteria, the
calculation of the FAV and CMC as shown in Table B, the Asiatic clam was
removed from the dataset to produce a criteria based on a “non-bivalve
mollusk” dataset (as commented on under question # 2 above).   There
may be water bodies with physical habitats (such as some headwater
streams) that are unsuitable for bivalve mollusks of any kind, not just
Unionids.   EPA should consider whether or not to base their
recommendations regarding separate criteria based on removing data for
Unionid genera only, or for all bivalve mollusks.  The same approach
should be used for both the acute and chronic criteria.       

(Note: there is a small typo in Table C; the far right hand column
should be labeled GMCV, rather than GMAV)

Use of 28-day Juvenile Test Data (see Appendix C):

Water quality criteria for the protection of aquatic life are derived
using toxicity endpoints that relate to population level impacts.  In
general, these endpoints relate to survival, growth and/or reproduction.
 The 28-day test with juvenile mussels, while similar in duration to a
standard chronic test, is not technically an early-life stage test
according to the 1985 Guidelines for Aquatic Life Criteria, as much of
the early development will have already occurred.

Question:

8.	Given that the juvenile life stage of freshwater mussels is
relatively long (2-6 years) are 28-day exposure tests with juvenile
mussels scientifically defensible as "chronic" test data for criteria
development?   

 Review Comment:

 	I do not believe that growth measured during a 28-day test can be
considered a true measure of a chronic effect as intended by the
requirements in the 1985 Guidelines for Aquatic Life Criteria, such that
they can be related to significant population impacts. Given the life
span of the organism and the time needed to reach maturity, a 20%
difference in growth observed over 28 days may be reversible and may not
be relevant on a population level and should not be used to derive
criteria.  

Question:

9.	Should toxicity studies on the growth rates of mussel shells during
28-day tests be considered quantitatively when developing water quality
criteria? 

Review Comment:

	No.  I do not believe that differences in growth rates measured over a
28-day period should be considered to provide acceptable quantitative
chronic endpoint data for freshwater mussels that should be used to
calculate chronic criteria for ammonia. Chronic criteria for ammonia
must be based on clearly recognized adverse effects in order to be
acceptable.  Reductions in survival are irreversible and are clearly
adverse effects, while slight reductions in growth may not be
significant to the population, or the organisms may be able to
compensate over time and reverse the difference in growth.   These type
tests do not meet the requirements of the EPA Guidelines for use as a
quantitative value in driving a chronic criterion.   Using such data to
derive chronic criteria, especially if it significantly influences the
final criteria would likely increase uncertainty about the final
criteria. 

Question:

10.	 Regarding the position statement and rationale on use of juvenile
mussel growth data – 

Is it scientifically defensible to include the juvenile growth data from
a 28-day exposure period as “other data” that might influence the
criteria however not be used directly in the derivation of the criteria
value? 

Review Comment:

Yes.   Although too many questions surround the use of growth effects
during a 28 day test in relation to long lived species such as
freshwater mussels to consider using differences in growth  as data that
directly influence the calculation of the final chronic criteria, these
data still deserve consideration and inclusion in the EPA criteria
document.  These data do provide useful information to help judge the
potential for “other” effects that have a potential for concern.
Including such data under the category of “other data” can provide
additional information that can lend confidence that the criteria based
on survival can provide an adequate amount of protection.  The magnitude
of difference in growth may make a difference; a 20 to 25 % reduction in
growth may not be of much ecological significance, especially if there
is a thought that the effect may be short term and reversible, while a >
75% reduction in growth may be considered more of a concern.  The
magnitude of difference between the effects on survival and the effect
on growth can also provide useful information.  Altogether, I see value
in presenting these growth data under the classification of “other
data” in the criteria document.

Question

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

Review Comment:

I would agree that this statement could also be true.  We simply do not
know whether or not an effect on growth observed over a 28-day exposure
can be considered reversible or indicative of a permanent effect.  As
discussed above, I believe the magnitude in difference in the growth and
the magnitude of difference between the survival end point verses the
growth endpoint could influence the level of concern regarding such
data.  

Question:

The values of the acute and chronic ammonia criteria have a strong
dependence on pH.  Juvenile and adult mussels, as sediment-dwelling
organisms, inhabit a medium that may have vertical pH gradients, thereby
creating some uncertainty about the appropriate pH to assign as their
exposure conditions.  For applying a criterion protecting mussels, do
you have suggestions on how states and EPA might determine the pH
applicable to the sediment micro-environment to which mussels are
typically exposed?

Review Comment:

The toxicity data currently available for freshwater mussels are for
glochidia and juveniles.  The exposure to unencysted glochidia will be
primarily via water exposure, and once encysted on a host fish they will
be in a less exposed condition.  Juvenile mussels will also be initially
exposed to ammonia in the water column following dropping off
(excystment) from the host fish.  As the juvenile mussels are deposited
onto the substrate they may remain on the surface or may become situated
within a few centimeters of the surface. They will need to maintain a
position on or near the surface in order to ensure access to
well-oxygenated water as well as filterable food.  In all cases, the
exposure to ammonia will be over the gills via water; either surface
water or interstitial water and the water quality criteria can be
adjusted to account for any differences in pH between the water column,
the surface-sediment interface or pore water at a specific depth in the
sediment.  

For normal use in establishing permit limits, or assessment of
waterbodies; the criteria should be applied to the water column.  If
there is a special concern at a site, the option of monitoring the pH of
interstitial water can be investigated, and the pH characteristic of the
interstitial water can be used to assess the criteria.  Care should be
taken to determine the pH in the interstitial water at depths where
mussels are likely to be found at the site.



Question:

12.	In general, should the criteria include a consideration for the
potential pH difference between sediment and the water?  If so, what is
the most scientifically defensible way to account for these differences
when deriving protective water quality criteria?

Review Comment:

The potential differences in pH between water and sediment is not so
much directly related to the criteria development, as it is an issue
related to how and where the criteria may be applied.  The water quality
criteria should provide recommendations for protective criteria based on
ammonia concentrations in water and adjusted for pH and temperature.  On
a site-specific basis, if in sediment there are significant differences
between pH in the interstitial pore water in sediment and the overlying
surface water,  this can be assessed by site-specific measurement of pH
levels in the sediment (at depths where juvenile mussels are likely to
be found) and then adjusting the criteria to those pH conditions. 
Another concern is whether or not the pH at the appropriate depth is
relatively constant.  These issues will likely vary considerably
depending on the characteristics of the sediment, and perhaps
seasonally.  The criteria can be adjusted to the pH in the interstitial
water to assess the potential for adverse effects at the site.

 

Question:

13.	Should exposure tests on juvenile mussels be conducted with or
without sediment in the test chamber?  

Review Comment:

Results of toxicity tests without sediment provide good information
regarding the toxicity of ammonia via exposure to ammonia concentrations
in water.  Acceptable control mortality in a test remains a significant
indicator that the test conditions and the general overall health of the
tested organisms are reasonable and the increasing mortality observed in
the test chambers with higher concentrations of the tested substances is
a direct result of the toxicity of the toxic substance.  If the
mortality in the control chambers remains low then, this is good
evidence that the tested organisms are not experiencing unacceptable
stress during the test.  In several water-only 28-day tests, 100% to 90
% survival was observed in controls. This provides evidence that
water-only tests with juvenile mussels can provide acceptable conditions
for assessing the toxicity of a contaminant.  

Water-only exposure tests provide relatively comparable and reproducible
exposure conditions that are useful for assessing toxicity sensitivity
and to distinguish effects of pH temperature or hardness etc.  Toxicity
testing in sediment will not change what we know about the sensitivity
of the species to ammonia via a water-only exposure as used and
summarized in this criteria document. As shown in Tables 1, 2 and 4, the
variability of toxicity values within a species and within genera is all
within a factor of 10.  Water-only toxicity tests methods have shown
good intra and inter laboratory variability for toxicity tests with
ammonia, as well as copper and chlorine.  This lends confidence in these
data for use in criteria development.  Toxicity data from water-only
exposures are most relevant to the purposes that water quality criteria
are most often employed; assessments of natural waters by comparison to
the criteria and for use in developing regulatory limits for permitted
discharges of the pollutant.  

Concerns that tests conducted in sediment might produce different
toxicity results are speculation.  Some speculation is that water-only
exposure stresses the test organisms and may make the test organisms
more sensitive to the toxicant.  However the good survival in the
control chambers indicates that the tested mussels are not overly
stressed during the test.  Also, test conditions are designed to provide
optimum dissolved oxygen and temperature conditions.  Natural conditions
often are less than optimum with lowered oxygen levels and other
conditions.  The sediment may be a more natural environment for the
juvenile mussels, making them more hearty and resistant to the toxic
substance, or the sediment may act to absorb the toxic substance and
make it less bioavailable and less toxic.  On the other hand, if the
sediment may absorb the toxic substance and over time increase the
concentration in the interstitial pore water in relation to the
overlying water.   Most of these issues are site specific in nature and
beyond the scope of national recommended criteria. 

Natural sediments are highly variable and even a standardized
reconstituted “clean” sediment to be used in laboratory tests may
prove variable in some unknown way perhaps in response to some
unrecognized factor, however this is simply speculation.  Use of
sediment in toxicity tests will likely introduce added variability
without providing additional value for judging sensitivity and this will
introduce added complexity to interpreting the tests’ results for
criteria development.  

Several studies (Whiteman et. al., 1996, Hickey and Martin, 1999) have
reported that ammonia in interstitial pore water can often be higher
than concentrations in the water.     Overall, these concerns about
increased ammonia concentrations in interstitial pore water are not so
much related to toxicity sensitivity or issues that directly influence
the derivation of criteria.  These concerns and observations are more
related to an exposure assessment of potential risk to the sediment
dwelling organisms and the related physical factors that influence the
exposure conditions.  

I believe that toxicity data from water-only tests that were conducted
under acceptable conditions (with acceptable control mortality) provide
the most useful data to judge the quantitative toxic effects of the
toxic substance.   These water-only data can then be used to assess
potential for toxicity in the water column.  Currently, there is
insufficient information available to indicate that the toxicity of
ammonia should be considered fundamentally different in interstitial
pore water as compared to the toxicity observed in water-only tests
under the same pH and temperature conditions.  An ammonia toxicity test
with water-only exposures provides reliable test results that can be
applied to exposures if the juvenile mussels are exposed to
concentrations of ammonia in water.    Water exposure tests should be
used in the derivation of water quality criteria and this information
should apply equally well to water column or interstitial pore water. 

If there is a special need to assess interstitial pore water at a site,
the criteria can be applied to an assessment of the interstitial pore
water.  Samples of interstitial water can be collected and analyzed for
ammonia concentrations and pH can be measured in the field.  If
monitoring of interstitial pore water is conducted, care should be taken
concerning the depth at which the samples are taken.  Juvenile mussels
are likely to be within a few centimeters of the surface and juveniles
and adult mussels are capable of some control over maintaining their
position in the sediment and should not be assumed to be regularly found
buried in deeper sediments where low dissolved oxygen, reduced food
availability and other stress factors may overshadow risk due to ammonia
toxicity.  Water samples for ammonia analysis and pH field measurements
should be taken at the surface interface or at a depth where mussels are
likely to be found. 



PEER REVIEW COMMENTS FROM

Steven P. Canton. MS

Senior Aquatic Ecologist

GEI Consultants

Littleton, CO

CHARGE QUESTIONS

Acute criteria in fresh waters:

1.	Are the toxicity tests used to derive the criteria scientifically
defensible for such use?  Are you aware of other relevant data that were
not used?

Yes, in general the toxicity tests used to derive the water column
criteria, as cited in the draft update, are scientifically defensible. 
Plus, it appears they provide sufficient data to evaluate inter and
intra-species comparisons using the guidelines established for deriving
water quality criteria.  

Additionally, a few years ago, we conducted a review of a number of
ambient water quality criteria (AWQC) as part of the Arid West Water
Quality Research Project (AWWQRP 2006).  One aspect of that project
included a review and update to the 1999 EPA Ammonia criteria document,
similar to this draft.  Checking those updates side-by-side, it appears
there are a few more studies that could be included in this
re-evaluation:

Augspurger, T., A.E. Keller, M.C. Black, W.C. Cope, and F.J. Dwyer.
2003. Water quality guidance for protection of freshwater mussels
(Unionidae) from ammonia exposure. Environmental Toxicology and
Chemistry 22:2569-2575.

Khatami, S. H., D. Pascoe, M. A. Learner.  1998.  The acute toxicity of
phenol and unionized ammonia, separately and together, to the
ephemeropteran Baetis rhodani (Pictet). Environmental Pollution
99:379-387.

Rubin, A.J., and M.A. Elmaraghy. 1977. Studies on the toxicity of
ammonia, nitrate and their mixtures to guppy fry. Water Research
11:927-935.

Russo, R.C., D.J. Randall, and R.V. Thurston. 1988. Ammonia toxicity and
metabolism in fishes. In R.C. Ryans, ed. Protection of River Basins,
Lakes, and Estuaries. pp. 159-173. American Fisheries Society, Bethesda,
Maryland.

Sangli, A. B. and V. V. Kanabur. 2001. Toxicity of ammonia to a
freshwater fish, Gambusia affinis and its effect on oxygen consumption.
Geobios 28:56-58.

Tomasso, J.R., C.A. Goudie, B.A. Simco, and K.B. Davis. 1980. Effects of
environmental pH and calcium on ammonia toxicity in channel catfish.
Transactions of the American Fisheries Society 109:229-234.

I should also note that in that prior review of the 1999 Ammonia
document (AWWQRP 2006), we raised concerns regarding the appropriateness
of some studies/data used in the 1999 EPA document (see Table 1 below). 
These studies may need an additional evaluation to determine if they
provide toxicity data meeting EPA data quality guidelines (Stephan et
al. 1985). 



Table 1

Deletions of Inappropriate Data Presented in the EPA 1999 AWQC Ammonia
Document (from AWWQRP 2006)

Species	References	Comments

Acute

Simocephalus vetulus	Mount 1982	Insufficient data to validate results

Oncorhynchus mykiss	Calamari et al. 1977	Unable to validate citation or
data-other data available by same author with this species

Oncorhynchus mykiss	Thurston et al. 1981	Unable to validate citation or
data

Oncorhynchus mykiss	Reinbold & Pescitelli 1982	Unable to validate
citation or data

Pimephales promelas	Thurston et al. 1981	Unable to validate citation or
data

Pimephales promelas	Reinbold & Pescitelli 1982	Unable to validate
citation or data

Gambusia affinis	Wallen et al. 1957	Insufficient data to validate
results

Lepomis macrochirus	Reinbold & Pescitelli 1982	Unable to validate
citation or data

Sander vitreum	Reinbold & Pescitelli 1982	Unable to validate citation or
data

Chronic

Ictalurus punctatus	Colt & Tchobanoglous 1978	Insufficient data to
validate results

Source: (AWWQRP 2006)





What are the technical considerations that EPA should evaluate when
mussels are present and mussels are absent with respect to the
recommended acute criteria?

Interestingly, we made a similar recommendation for this particular
situation (i.e., the concept of with and without mussels-based criteria
for ammonia) in that prior report (AWWQRP 2006, Chapter 4).   I have
attached a copy of the relevant chapter from that analysis, which may be
useful, as least in reference to how this has been looked at by other
parties. 

In addition to a with and without mussel approach, the document could
also include other “implementation” issues, such as evaluating
additional site-specific information needed to provide in-depth
technical recommendations.  For example:

Is there a definitive cause and effect relationship between ammonia
sources and the diversity and abundance of mussels downstream of the
source?

How have the manageable sources (i.e., Agriculture, WWTP) of nitrogen
changed over the last decade within the study waters in comparison to
measured changes in mussel communities?  

Is there evidence of anthropogenic nitrogen loading that is equal to or
greater than the natural occurrence of decomposition due to
heterotrophic bacteria?

What is the within-site spatial and temporal variability in NH3 of the
pore water and is this variability linked to season, storm events,
temperature, or flow conditions.

In addition it would be advisable to include a statement to recognize
the importance of developing “site-specific standards”, since
mussels now play an important role and are the most sensitive species in
the list. Site-specific standards reflect the reality of which species
are needed to be protected in a particular waterbody, particularly with
mussels that may be more abundant in some parts of the country, whereas
in some other regions their presence can be very limited.  In this case,
some guidance on when and how to do mussel surveys to determine when to
apply a “with mussels” or “without mussels” criterion would be a
useful addition to the criteria document.

3.	Is it scientifically defensible to exclude the glochidia data at this
time due to the uncertainty of appropriate test duration time for this
life stage?  Do you believe there is an alternative approach to the use
of this data that would be more scientifically sound?

Given the rationale provided, in combination with EPA criteria
development guidance, general ecotoxicological principles, and my own
views on general invertebrate biology, I agree it is scientifically
defensible to exclude glochidia data in ammonia criteria development for
the following reasons. 

First, the document notes that despite having  four acute toxicity
studies that consider five different mussel species (with the glochidia
stage) ,with the majority of test having a duration of 2 days, there are
still several questions to be answered when testing these particular
life stages:

How long do they stay attached to the host?

How long do they stay in conglomerates and does this jelly (for those
species) protect them from exposure to contaminants?

If they sink to sediment how long do they stay there in the conglomerate
before transforming to the juvenile mussel?  

I agree with the authors that these valid questions and also note that
the questions will probably have different answers for different species
of mussels, which makes the data of even more uncertain general
ecological relevant and, therefore, use of glochidia data even more
problematic and difficult to apply generally to this group for criteria
development purposes.

 Second, concerning the mechanisms of ammonia toxicity in this
particular stage of development there is also a long list of unanswered
questions: would glochidia be able to take up ammonia if available in
the host?, what is the mechanism of glodichia exposure to ammonia when
they are released in conglomerates?  - Again – great uncertainty with
unknown ecotoxicological significance.

Third, based on the results from those four studies it seems that if the
glochidia acute data were included in the calculations, they would
dramatically reduce the acute criteria and probably be over conservative
for other fresh water species without actually knowing if the
information has any ecological relevance.   

Given these uncertainties – which are well articulated in the draft
document, there appear to be two scientifically defensible approaches:
1) Do not use the glochidia acute toxicity data 2) Do a final acute
criteria calculation “with and without glochidia”. The first
approach would be the easiest to avoid confusion and possible errors in
development of safe levels for ammonia.  The second approach would give
flexibility – although it would be necessary to include a strong
caveat to a “with glochidia” value, given the questions noted in the
criteria document and my thoughts above. 

4.	Regarding the proposed approach to glochidia data in the 2009 draft
position statement as it relates to ecological relevance and
practicality - Is the approach a scientifically defensible principle for
structuring the population exposure duration problem and designing
further research to quantify such duration?

The proposed approach will give researchers and scientific community
time to come up with some answers and provide relevant information
concerning the appropriate time of duration for glochidia exposure
testing – if such testing is determined to be necessary for this
short-lived, parasitic life stage.  However, if this approach is
proposed, it will also create a limbo period, in which this issue will
have to be addressed while decisions are made for application of a
criteria that some may fight as “non-protective” since the criteria
document has “acknowledged” that a sensitive life stage was not
considered!

I’ll be the first to admit it is a difficult to task to structure a
criterion that includes population exposure approach to criteria
development when dealing with a wide variety of fresh water mussel
species that have different life spans (few months to several years),
not to mention a glochidia life-stage that can be released individually
or in conglomerates, and that can get attached to fish, rocks, sediment
or plants. 

Such variability could mean that for the particular issue of including
glochidia data or not, we perhaps should be treating these mussels
individually, by genus or species, rather than thinking of them simply
as the “mussel” group.  Regardless, it will take a considerable
amount of time and effort to develop some way to make “uniform”
relevant exposure toxicity testing time for the glochidia life stage
that will be relevant for all if not the majority of fresh water
mussels.

Hyalella azteca position statement and proposed rationale (see Appendix
B):

5.	Are the position statement and supporting rationale regarding use of
toxicity data for Hyalella azteca in criteria development reasonable and
scientifically sound recommendations?  

I support the position statement based on the workgroup’s review of a
number of toxicity tests on Hyalella, referenced in the rationale.  I
can’t help but wonder if this recommendation holds for Hyalella test
data for other EPA criteria documents, as well – for example, Hyalella
is the most sensitive species in EPA’s chronic cadmium database (EPA
2001).  Is this position statement generic to all EPA criteria
databases, or simply this ammonia update?  

I should note that our prior review of the study also recommended
removal of Hyalella from the ammonia database, primarily because of the
poor control performance (AWWQRP, 2006), which it now appears may be due
to the ionic balance issue noted by the position statement.  So, I
support non-use of Hyalella toxicity data for this ammonia update.

Most of the recommendations provided have scientific basis. It appears
that ongoing studies will provide quality data that will widen the
knowledge for water quality requirements for H. azteca husbandry.
However, the statement concerning the complications due to possible H.
azteca genetic or taxonomic diversity is not relevant because the way
this information is used in criteria development, which only considers
appropriate data at genus level (i.e., GMAVs) not new species or
subspecies. In addition, the issue of a wide range of surface waters
inhabited by H. azteca could be easily addressed setting a site-specific
standard, which will consider only the species present at a particular
site.  Thus, inclusion of this information may add confusion rather than
help explain a decision on non-use for criteria development.

Chronic Freshwater Criterion:

6.	Are the toxicity tests and other studies used to derive the criterion
scientifically defensible for such use?  Are you aware of other relevant
data that were not used?

While, the toxicity tests used to derive the criteria are generally
scientifically defensible (see discussion below); I am uncomfortable
with the calculation because there is insufficient data to evaluate
intra-species comparisons using the guidelines established for deriving
water quality criteria.  More specifically, the 2009 data set still does
not fully meet the EPA guidelines for developing water quality criteria.
 A salmonid representative has been added to the data set (which was
missing in the 1999 document), but a representative from the Class
Insecta is still missing.  It appears that the option for using
acute-to-chronic ratios for determining chronic criteria would be more
relevant than simply using a “hypothetical GMAV for insects”.

One of the studies noted in Question 1 may provide additional relevant
chronic data.

It is also important to point out that there are only two studies
(Anderson et al. 1978, Sparks and Sandusky 1981) available for Musculium
genus. Interestingly, those two studies provide substantially different
estimates of what level of ammonia is protective to fingernail clams –
creating uncertainty in the true chronic value.  A closer analysis of
these two papers reveals that while the later Sparks and Sandusky study
was designed to confirm Anderson et al’s findings, the researchers
faced some serious methodology deficiencies, as the authors recognize:
“Although it would have been desirable to perfect the culture methods
first and then employ them in the toxicity tests, methods were developed
as testing proceed because of limited time” (Sparks and Sandusky
1981). As part of their study design, Sparks and Sandusky used
clinoptilolite to remove ammonia, but note this compound could also have
remove potassium and chlorinated hydrocarbons – which raises the
question that toxic effects observed in Musculium transversum in this
particular study may be also attributed to other confounding factors in
the test water and not exclusively to ammonia exposure. In addition,
authors conclude that also biological processes (i.e., naturally
occurring bacteria that converts ammonia into a relatively non-toxic
nitrate) may be involved in ammonia removal rather than physico-chemical
processes. Thus, it would be advisable to re-analyze the relevance of
these studies for this update.  As it stands, I cannot fully support
calculating a GMAV for this fingernail clam using these two studies.

Another issue that is relevant to mention is the continued inclusion of
a temperature relationship for chronic ammonia criteria originally used
by EPA (1999), which was derived from a single study. It appears the EPA
used the Arthur et al. (1987) study, which evaluated acute ammonia
toxicity to 14 species (9 invertebrate and 5 fish species), to
incorporate temperature dependence into the chronic equations. There are
three problems with incorporating temperature in the chronic
relationship, but not acute. First, there is lack of chronic ammonia
toxicity studies for fish or invertebrate species to demonstrate the
relationship between temperature and chronic ammonia toxicity. Second,
the rationale for deriving an invertebrate chronic temperature slope
from acute data (i.e., fish and invertebrates) is unclear. Finally, the
assumption that chronically exposed invertebrates will have similar
temperature dependence compared to acutely exposed individuals is based
on the Thurston et al. (1984) study and those authors indicated that a
96-h period is insufficient to determine an acute toxic concentration of
ammonia for insects. Consequently, it appears there are no appropriate
data to support incorporation of a temperature component to the chronic
ammonia standards, just as EPA concluded with the acute equations. The
position statement notes that the temperature-dependent toxicity model
now only applies to invertebrates. 

I have to admit I was hoping the actual equations, inflection points,
and pH and temperature relationships would be thoroughly re-evaluated as
part of this update, given the abundance of new data included not in the
1999 document.  I would strongly recommend re-evaluation of the chronic
temperature component of the equations rather than just accept the 1999
update evaluation.  Although, I have to admit, the draft document
doesn’t actually provide any equations to test – just recalculations
of acute and chronic endpoints.  This is another disappointment.

7.	Is the freshwater chronic criterion scientifically defensible with
mussels present and mussels absent?  

The freshwater chronic criterion based on a mussels present/absent
approach maybe scientifically defensible – but will potentially result
in very different approaches, even in regions of the United States where
mussels are known to be absent or rare.  In fact, the primary question
will become “what level of data is required” to determine whether
the habitat is or is not suitable to support mussels and/or whether
mussels have ever occurred within the site historically.  Guidance on
such data needs questions would be useful to include in the criteria
document to support a with/without mussel criterion.  

Additionally, even when the “mussel present” option is applied, a
whole other series of issues arise, because there is often a poor
correlation between the water column NH3 concentration and the pore
water NH3 concentration where most juvenile mussels reside. 
Site-specific sediment conditions can greatly influence the availability
of NH3, and in the case of the upper Mississippi River, as much as 6-30
times greater than the surface waters (Bartsch et al. 2003).  

In such cases, criteria developed without considering the complete
affects of sediment on the accumulation of NH3, or other confounding
affects on mussel communities (i.e., the fact that mussel beds can
self-produce nutrients, such as ammonia) (Dankers & Koelemaij 1989), may
greatly underestimate, or even overestimate, a criterion that is
protective of mussels.  In addition, this also brings up the question of
whether the use of sediment is appropriate in chronic toxicity tests
designed to be protective of aquatic life in the water column, and what
endpoint should be evaluated using such methodologies.  

Addressing the “Use of 28-day Juvenile Test Data” (see Appendix C):

8.	Given that the juvenile life stage of freshwater mussels is
relatively long (2-6 years) are 28-day exposure tests with juvenile
mussels scientifically defensible as "chronic" test data for criteria
development?   

Water quality criteria for the protection of aquatic life are derived
using toxicity endpoints that relate to population level impacts.  In
general, these endpoints relate to survival, growth and/or reproduction.
 I would agree that the 28-day test with juvenile mussels, while similar
in duration to a standard chronic test for many other groups of
organisms, is not technically an early-life stage test according to the
1985 Guidelines for Aquatic Life Criteria, and as such, should perhaps
not be included in the chronic database.

If this is carried forward, I would support the approach of the 28-day
exposure tests being considered as “other data’ – if the test are
acceptable in all other respects.  Reasons for not including in chronic
criteria calculations could follow and build on the arguments presented
in the position paper:  

First, it is not representative.  If the duration of the life stage is
2-6 years for juvenile life stage of fresh mussels, 10% of this amount
of time would be 73-219 days. Thus, a longer exposure (i.e., 90-day)
would be more representative for a “chronic” study for juvenile
fresh water mussels than 28-day tests. 

Second, extending the exposure time would allow the tested animals to
adapt better to laboratory conditions, thus reducing the stress and
decrease the risk of having test condition-biased results. 

Third, although it will be more challenging to keep all the water
quality parameters under control in a longer exposure it could also
provide other avenues for future research on juvenile mussels (i.e., if
captivity conditions and pollutant exposure would enhance growth over
longer periods of time, if they may need additional food source, among
others)

Once possible conflicting issue is that other species used in EPA
criteria development have “allowable” chronic test conditions that
also represent a small portion of their total life span – e.g., 60-day
post hatch ELS for salmonids, which can live for years.  The difference,
conceptually, is that the 28-day test for mussels does not include the
earliest life stage prior to juvenile testing commences – i.e., it
does not start with egg, then hatch, then juvenile exposure more typical
of an ELS for fish.  If we can conceptually overcome this “weakness”
for a 28-day juvenile mussel test, then I suppose growth-effects data in
a 28-day juvenile mussel test could be at least evaluated in the context
of other reported data – and if comparable to effects data from
longer-duration testing, then perhaps it could be used in some form for
criteria calculations.

9.	Should toxicity studies on the growth rates of mussel shells during
28-day tests be considered quantitatively when developing water quality
criteria?

As described in the previous question, the studies on growth rates of
mussels shells during 28-day tests should only be used as information to
better evaluate appropriate longer-term chronic studies (i.e., 90-day
studies). The results from longer chronic tests could certainly be used
quantitatively for the developing the water quality criteria. Unless
28-day test data provide similar endpoints to long-term testing, it
would be necessary to keep the results separate and potentially have two
conflicting sets of data (i.e., 28-day and 90-day data) – which will
only make the calculations of the new criteria more complicated, and can
lead to confusion when trying to implement.  And I have to say that
ammonia criteria development is confusing enough as is.

10.	 Regarding the position statement and rationale on use of juvenile
mussel growth data – 

Is it scientifically defensible to include the juvenile growth data from
a 28-day exposure period as “other data” that might influence the
criteria however not be used directly in the derivation of the criteria
value? 

My understanding of data included as “other data” in the criteria
document is that these data are provided just as a reference to make the
interested parties know that these data exist.  They can certainly be
presented that way and not considered in the calculations of the chronic
criteria for the reasons outlined in the previous two questions. 



Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

As noted in the position paper, it is not possible with the available
data to conclude that impaired growth after 28-day exposure will
necessarily lead to continued reductions in growth over time.  The
position paper rightly points out the possibility of organisms to
“recover” from potentially short term stress and show no further
decline. In fact, these organisms may exhibit normal growth during a
more relevant long-term test.  However, given the lack of information
available to compare short-term and long-term toxicity studies in fresh
water mussels, it would be advisable to note, as appropriate, that any
of the three outcomes: continued decline, stabilize, or recovery is
possible. In addition, the propensity for any of these three outcomes
will most likely depend upon the particular mussel species and the stage
of development used in the toxicity tests.

11.	The values of the acute and chronic ammonia criteria have a strong
dependence on pH.  Juvenile and adult mussels, as sediment-dwelling
organisms, inhabit a medium that may have vertical pH gradients, thereby
creating some uncertainty about the appropriate pH to assign as their
exposure conditions.  For applying a criterion protecting mussels, do
you have suggestions on how states and EPA might determine the pH
applicable to the sediment micro-environment to which mussels are
typically exposed?

Given that juvenile mussels of most species reside completely in the
sediment, filter pore water, and feed on the sediment, I would suggest
one approach might be determining a mean burial depth for the juveniles
of the selected mussel species within the 2009 data set.  In laboratory
tests, juvenile Villosa iris are reported to bury <1cm, yet were not
exposed to the overlying water (Yeager et al. 1994).  Perhaps it would
be useful to use this burial depth and evaluate the average pH
conditions within a burial zone, per se.  In addition, it is likely any
measures of pH in the mussel micro-environment (i.e., taking measures of
water quality over a mussel bed) may also show elevated ammonia levels
simply from the metabolic activity of those organisms, themselves.  I
can’t help but wonder how to determine attainment of “safe”
ammonia levels in future field monitoring, given the potential for
organism-generated ammonia values.

12.	In general, should the criteria include a consideration for the
potential pH difference between sediment and the water?  If so, what is
the most scientifically defensible way to account for these differences
when deriving protective water quality criteria?

Because adult mussels siphon directly from the water column, there is no
need for a pH differential calculation.  However, to be protective of
the juvenile life stage which typically filters pore water, there may be
a need for differential pH measurements.  But how does one account for
all of the biotic and abiotic factors that affect pH in the sediment? 
Because site-specific sediment conditions greatly influence the pH
levels, adjusting the pH should only be an option for site-specific
calculations (or considered as “other data”) rather than within the
national water quality criteria.

13.	Should exposure tests on juvenile mussels be conducted with or
without sediment in the test chamber?  

If the test is designed to address water column concerns, then the
exposure test using juvenile mussels should be conducted without
sediment in the test chamber.  The use of sediment in water quality
toxicity tests introduces too many confounding factors in tests designed
to evaluate the needs to be protective of aquatic life in the water
column.  For many species of mussels, the juveniles primarily feed on
particles in the sediment and filter the pore water, rather than
directly siphon from the overlying water column.  Depending upon the
composition of the sediment (i.e., organic matter content) the sediment
layer can have profound influences on the pH and ammonia concentration
within the pore water of the sediment.  In addition, the metabolic
byproducts of juvenile organisms that live in sediments also influence
the level of ammonia observed within pore waters of sediment.  For these
reasons alone it is difficult to separate the effects solely due to
conditions within the water column versus the effects of the sediment
layer.

Furthermore, in static renewal tests that incorporate the use of
sediment, there is the potential for decreased ammonia concentrations,
as well as pH, in the pore water of sediment when compared to the water
column conditions (Allan and Maguire 1992); whereas in natural
conditions this relationship is often reversed, such that pH levels and
ammonia concentrations are greater in the sediments than the overlying
water column.   The sediment based toxicity studies of Whiteman et al.
(1996) tried to address these gradient issues between the sediment and
water column, by introducing nitrogen via sediment and maintaining
overlying water concentrations in a non-toxic range.  However, the
mobility of the test organisms allowed them to seek out the zones of
lower ammonia concentration at the interface rather than within the
sediment, which also confounded the test results.

This concludes my responses to the Charge Questions.  Given the time
constraints, I feel somewhat limited in what I could provide.  I hope
this review was helpful – and thank you for the opportunity.

References:

Allan, G. L. and G. B. Maguire. 1992. Effect of pH and salinity on
survival, growth and osmoregulation in Penaeus monodon Fabricius.
Aquaculture 107:33-47.

Anderson, K.B., R.E. Sparks, and A.A. Paparo. 1978. Rapid Assessment of
Water Quality, Using the Fingernail Clam, Musculium transversum. WRC
Research Report No. 133. University of Illinois, Water Resources Center,
Urbana, IL.

Arid West Water Quality Research Project (AWWQRP).  2006.  Evaluation of
EPA Recalculation Procedure in Arid West Effluent-Dependent Waters:Final
Report. Prepared for the Arid West Water Quality Research Project by URS
Corporation, Chadwick Ecological Consultants, Inc., and Parametrix,
Inc., Pima County Wastewater Management Department, Tucson, AZ.

Arthur, J.W., C.W. West, K.N. Allen, and S.F. Hedtke. 1987. Seasonal
toxicity of ammonia to five fish and nine invertebrate species. Bull.
Environ. Contam. Toxicol. 38:324-331.

Augspurger, T., A.E. Keller, M.C. Black, W.C. Cope, and F.J. Dwyer.
2003. Water quality guidance for protection of freshwater mussels
(Unionidae) from ammonia exposure. Environmental Toxicology and
Chemistry 22:2569-2575.

Bartsch, M.R. et al. 2003. Effects of pore water ammonia on in situ
survival and growth of juvenile mussels (Lampsilis cardium) in the St.
CroixRiverway. Environ Toxicol Chem 22:2561-2568

Calamari, D., R. Marchetti, and G. Vailati. 1977. Effects of prolonged
treatments with ammonia on stages of development of Salmo gairdneri.
(English Translation) Nuovi Annali d’Igiene e Microbiologia.
28:333-345.

Colt, J.E., and G. Tchobanoglous. 1978. Chronic exposure of channel
catfish, Ictalurus punctatus, to ammonia: effects on growth and
survival. Aquaculture 15:353-372.

Dankers N, K. Koelemaij. 1989. Variations in the mussel population of
the Dutch Wadden Sea in relation to monitoring of other ecological
parameters.  HYPERLINK
"http://www.springerlink.com/content/103796/?p=265c1a04a27c42bdaafba7b3b
6b9a9b9&pi=0" Helgoland Marine Research   HYPERLINK
"http://www.springerlink.com/content/t07111844315/?p=265c1a04a27c42bdaaf
ba7b3b6b9a9b9&pi=0" 43:  529-535

Khatami, S. H., D. Pascoe, M. A. Learner.  1998.  The acute toxicity of
phenol and unionized ammonia, separately and together, to the
ephemeropteran Baetis rhodani (Pictet). Environmental Pollution
99:379-387.

Mount, D.I. 1982. Ammonia toxicity tests with Ceriodaphnia acanthine and
Simocephalus vetulus. Memorandum to Dr. R.C. Russo, ESEPA, August 6,
1982.

Reinbold, K.A., and S.M. Pescitelli. 1982. Acute Toxicity of Ammonia to
the White Sucker. Final Report to EPA. Center for Aquatic Ecology,
Illinois Natural History Survey, Champaign, Illinois.

Rubin, A.J., and M.A. Elmaraghy. 1977. Studies on the toxicity of
ammonia, nitrate and their mixtures to guppy fry. Water Research
11:927-935.

Russo, R.C., D.J. Randall, and R.V. Thurston. 1988. Ammonia toxicity and
metabolism in fishes. In R.C. Ryans, ed. Protection of River Basins,
Lakes, and Estuaries. pp. 159-173. American Fisheries Society, Bethesda,
Maryland.

Sangli, A. B. and V. V. Kanabur. 2001. Toxicity of ammonia to a
freshwater fish, Gambusia affinis and its effect on oxygen consumption.
Geobios 28:56-58.

Sparks, R.E. and M.J. Sandusky. 1981.  Identification of Factors
Responsible for Decreased Production of Fish Food Organisms in the
Illinois and Mississippi Rivers.  Final Report for Project No. 3-291-R,
Illinois Natural History Survey, River Research Laboratory, Havana,
Illinois.

Stephan, C.E., D.I. Mount, D.J. Hansen, J.H. Gentile, G.A. Chapman, and
W.A. Brungs.  1985.  Guidelines for Deriving Numerical National Water
Quality Criteria for the Protection ofAquatic Organisms and Their Uses. 
PB-85-227049.  U.S. Environmental Protection Agency, Office of Research
and Development, Duluth, Minnesota.

Thurston, R.V., R.J. Luedtke, and R.C. Russo. 1984. Toxicity of Ammonia
to Freshwater Insects of Three Families. Technical Report No. 84-2,
Fisheries Bioassay Laboratory, Montana State University, Bozeman, MT. 26
pp.

Thurston, R.V., R.C. Russo, and G.A. Vinogradov. 1981. Ammonia Toxicity
to Fishes. Effect of pH on the Toxicity of the Un-ionized Ammonia
Species. Environ. Sci. Technol. 15:837-840.

Tomasso, J.R., C.A. Goudie, B.A. Simco, and K.B. Davis. 1980. Effects of
environmental pH and calcium on ammonia toxicity in channel catfish.
Transactions of the American Fisheries Society 109:229-234.

U.S. Environmental Protection Agency (USEPA). 1999 Update of Ambient
Water Quality Criterion for Ammonia. Office of Water, Washington, DC.
EPA-822-R-99-014.

U.S. Environmental Protection Agency (USEPA).  2001.  2001 Update of
Ambient Water Quality Criteria for Cadmium.  EPA-822-R-01-001.  Office
of Water, Washington, DC.

Yeager, M.M., D.S. Cherry and R.J. Neves. 1994. Feeding and burrowing
behavior of juvenile rainbow mussels, Villosa iris (Bivalvia:
Unionidae). J.N. Am. Benthol. Soc. 13(2):217-222.

Wallen, I.E., W.C. Greer and R. Lasater. 1957. Toxicity to Gambusia
affinis of certain pure chemicals in turbid waters. Sewage Ind. Wastes.
29: 695-711.

Whiteman, F.W., G.T. Ankely, M.D. Kahl, D.M. Rau, and M.D. Balcer. 1996.
Evaluation of interstitial porewater as a route of exposure for ammonia
in sediment tests with benthic macroinvertebrates. Environmental
Toxicology and Chemistry. 15(5):794 - 801.



PEER REVIEW COMMENTS FROM

Jerome M. Diamond, Ph.D.

Principle Aquatic Ecologist and Director of Ecotoxicology

Tetra Tech

Owning Mills, MD Peer Review Ammonia

Acute criteria in fresh waters:

1.	Are the toxicity tests used to derive the criteria scientifically
defensible for such use?  Are you aware of other relevant data that were
not used?

Response:  In general, the toxicity tests used to derive ammonia
criteria in the draft Reassessment Document (RD) are scientifically
defensible given EPA’s criteria development guidelines (Stephan et al.
1985).  However, I have reservations regarding the use of snail data
derived from field-collected organisms as explained below under Chronic
Freshwater Criterion.  I am not aware of other relevant freshwater
toxicity test data that were not either used or considered in the RD.

2.	What are the technical considerations that EPA should evaluate when
mussels are present and mussels are absent with respect to the
recommended acute criteria?

Response:  Attempting to differentiate freshwater ammonia criteria on
the basis of mussels being present or absent is in my opinion a very
tenuous and probably unsupportable idea scientifically.  First, the
acute data indicate that Corbicula (Asiatic Clam) is nearly as acutely
sensitive as the most sensitive unionids.  Corbicula is not a unionid
and has a totally different life cycle than unionids.  Furthermore, they
are now widespread in the U.S. so it may prove fairly inefficient to
separate criteria on the basis of presence of bivalves in general.  

As an implementation issue, distinguishing between surface waters with
mussels and those without is not nearly as straightforward as it is for
salmonids or “cold water” species presence/absence because the
latter are dependent on a fairly known, measurable water characteristic
(water temperature regime).  Also, while most states have cold water as
well as warm water aquatic life uses in their water quality standards,
I’m not aware of many (perhaps any) that have specifically a mussels
use.  The only example of which I’m aware is Virginia DEQ, which
distinguishes different application of the chlorine standard based on
whether federally listed mussels are known to be present in a given
stream segment. They apply a “halogen ban” (no chlorine or other
halogenated disinfection chemicals are allowed to be used for wastewater
treatment) in those segments for those permittees that do not have
intermittent dischargers and have discharges <  20,000 gallons per day. 
 However, in this case the water quality standard was not changed but
rather a risk management approach was taken to reduce the major point
sources of chlorine or other halogenated disinfection chemicals to the
stream.

Mussels can and do occur (or have occurred) in a wide variety of habitat
types and water quality regimes and so their presence is not often known
without specifically sampling for them.  Such sampling may need to be
fairly intensive to “prove” the absence of mussels.  In addition,
mussels have been extirpated from a number of streams in the U.S. due to
historic land use and habitat changes, toxic spills, and poorly treated
wastewater and nonpoint runoff (e.g., see EPA 2002).  If today, a state
finds that mussels are not present but historically they were, what then
should be the ammonia criteria?  I believe such differentiations in
criteria suggested in the RD be relegated to EPA’s recalculation of
site-specific criteria (USEPA 1994).  That procedure was designed to
identify whether certain types of species are present or could occur in
a given site or waterbody.  This would be a more efficient and
scientifically defensible way to handle the apparent sensitivity of
freshwater mussels to ammonia.

3.	Is it scientifically defensible to exclude the glochidia data at this
time due to the uncertainty of appropriate test duration time for this
life stage?  Do you believe there is an alternative approach to the use
of this data that would be more scientifically sound?

Response:  I agree with the RD that glochidia data should not be
included in criteria derivation at this time but not entirely for the
reasons given by EPA.  First, a female mussel often produces hundreds or
even thousands of glochidia in a single season.  This helps counteract
the fact that a high percentage of the glochidia will not survive
because they have not encysted on a host fish within a reasonable amount
of time.  As many researchers have documented, glochidia are incapable
of surviving and maturing on their own—they require nourishment from a
host organism typically within days of being produced.  It could be
argues that basing criteria on a 50% reduction in glochidia survival
would be similar to basing criteria on a 50% reduction in egg survival
of other species that broadcast an abundance of eggs (e.g., marine
barnacles) in which only a small fraction of the larvae are likely to
survive and become established under natural, unstressed conditions.  I
am not aware that EPA has ever based acute criteria on egg survival
(even fertilized egg survival) and I believe this might contradict
EPA’s criteria development guidelines for establishing acute criteria.

A second reason I don’t think glochidia should be included at this
time is that the majority of mussel species at risk today (i.e.,
threatened, endangered, or species of concern) appear to have relatively
narrow specificity in terms of a vertebrate host and have evolved
structures (e.g., lures) or behaviors that enable very rapid encysting
in the host (e.g., see Rogers-Lowery and Dimock 2006).  Of the mussel
species actually used in glochidia testing that are discussed in the RD,
most of them have evolved adaptations to enable rapid infestation of a
host (often < 6 hours), resulting in a very short free-living glochidia
stage.  Thus, for the majority of mussel species at risk, exposure
periods > 24h are probably inappropriate, regardless of whether control
survival is > 90%.

A third reason is that while the ASTM glochidia test procedure appears
to be reasonably robust in terms of intra and inter-laboratory testing,
the procedure is still a “Guide”, not a “Method”, which
indicates that there are still many aspects requiring expert judgment,
trial and error, and research, especially in terms of culturing and
maintaining glochidia  under laboratory conditions.  Also the inter-lab
study reported by Wang et al. (2007) used the same water source and
glochidia stocks for all labs.  This would tend to underestimate the
true interlab variability, where each lab uses their own source of
glochidia (and techniques for obtaining glochidia) as well as water
source.  In addition, at the present time there are still relatively few
laboratories that have performed this test (e.g., compare with daphnia
or amphipod tests).  Much of the glochidia data under consideration were
produced by a few laboratories.  This may be acceptable for tests that
use species closely resembling those for which a standard EPA or ASTM
method has been developed and used extensively; e.g., an acute test with
a different species of minnow or Daphnia but the identical test design
and organism life stage as used for the standard fathead minnow or
Daphnia acute test.  However, such is not the case with the ASTM
glochidia test, which uses a fairly different test design and the
endpoint is dependent on sensitivity to a sudden stress (i.e., rapid
valve closure to NaCl exposure).  While the ASTM procedure appears to
address these concerns for the most part, there is still some
uncertainty in my mind regarding defensibility of the test endpoint.  As
noted in the ASTM method, exposures beyond 24h may not achieve 90%
control survival, indicating the precarious nature of keeping this
lifestage alive without a host.  Given this, is the apparent increase in
ammonia sensitivity between 6 and 48h a true difference in sensitivity
or is it an artifact of the method (i.e., glochidia are under increased
stress).  I also question the ecological importance of having valve
closure within one minute or less to the introduction of a NaCl
solution.  I’m not altogether sure this endpoint is similar to other
acute endpoints relied upon by EPA in their water quality criteria.  In
addition, the ASTM Guide points out that live or dead mussels could be
open or closed at a given time and some may respond more slowly to the
NaCl shock than others (Kernaghan et al. 2005).  If there was a more
direct way to determine mortality of glochidia to ammonia exposure, for
example, that would be preferable in terms of using such data for
criteria development.

I wonder if better use of glochidia survival data couldn’t be made by
considering the level of juvenile recruitment necessary to maintain a
viable mussel population, similar to the EPA fish larval recruitment
model for dissolved oxygen in coastal waters in the Virginian Province
(USEPA 2000).  Perhaps a similar type of framework could be used based
on the mussel species (or, if feasible and defensible, genus or even
family), when glochidia are typically produced in a given region, and a
function specifying ammonia toxicity to glochidia given ambient
temperatures and pH at that time.  The modeling framework could be made
general so that a user could input necessary temperature and pH data,
recruitment timing (e.g., season, month) and certain assumptions
regarding host abundance.  This would not be unlike the larval
recruitment model EPA has developed for dissolved oxygen with the
exception that the abundance of another species (i.e., a host) is not
needed or relevant for the dissolved oxygen model.  However, if such a
model was considered (albeit simplistic but using conservative
assumptions based on expert malacologist judgment), it could put
glochidia toxicity testing results into an ecological context in a more
useful way.

4.	Regarding the proposed approach to glochidia data in the 2009 draft
position statement as it relates to ecological relevance and
practicality - Is the approach a scientifically defensible principle for
structuring the population exposure duration problem and designing
further research to quantify such a duration?

Response:  I am not sure I concur entirely with the proposed approach
for further research discussed in Appendix A of the RD.  I do agree that
it would be useful to have better information regarding the natural life
expectancy of free-living glochidia in nature for several representative
species covering the range of different reproduction strategies
observed.  Such information would help indicate whether a 6, 24, or 48h
exposure is warranted.  I do not, however, understand EPA’s proposal
specifying the free-living duration based on 95% of the glochidia that
attach to a host.  This appears to be a fairly restrictive proposition
and it is not clear to me how this would be determined.  As
Rogers-Lowery and Dimock (2006) observed, the encapsulation process in
fish varies with the species and organism history of exposure to mussel
infestations.  Many researchers have demonstrated poor encapsulation
rates of glochidia of certain mussel species (including several listed
as threatened or endangered) with several common fish species (e.g.,
Dodd et al. 2006).  In other words, the number of glochidia that encyst
on a host has as much to do with the host as it does glochidia viability
in and of itself.  Somehow, there would need to be the presence of a
known host with the glochidia to determine an answer to EPA’s
proposal.  However, viable host species are not known for many mussel
species.  Therefore, this evaluation would require testing with a few
mussel species for which host species are known with certainty.

Other comments regarding Appendix A

Under “USEPA’s Aquatic Life Criteria Coordinating Committee”
comments, point #2 makes it appear that EPA is willing to alter a
national criterion due to a commercially valuable species (e.g., trout)
but not an ecologically critical species?  Point #3:  I agree there is
no evidence that glochidia in conglutinates are exposed to water-borne
pollutants but there’s no evidence that they are not exposed either. 
Also, uptake of pollutants by planktonic invertebrates lacking gills
(e.g., Daphnia) occurs via absorption through the carapace, not just
through food.  Point #8:  “For glochidia that attach to hosts,” –
all glochidia are meant to attach to hosts.  What I think you mean here
is “In terms of the duration of free-living glochidia, there is a
major difference….”



Other comments regarding acute data in the Reassessment Document

Daphnia GMAV is Incorrect

I believe the document reports an incorrect GMAV for Daphnia (p. 9) as
well as test data in Table 1 (p. 49-50).  It appears that the unionized
ammonia values reported by Russo et al. (1985) and reported as such in
EPA’s 1999 document, were incorrectly transcribed as total ammonia
values (see p. 109, Appendix 4 of 1999 document as well as p. 38, Table
4 of the 1999 document).  Even with the temperature correction for
invertebrates in the Reassessment Document, Daphnia couldn’t possibly
go from Rank 19 in acute sensitivity to 8th in the draft document.  This
error has other ramifications as well:  assuming the Daphnia value is
incorrect in the draft, when mussels were absent, the four most
sensitive genera are Corbicula, Potamopyrous, Pleurocera, and Prosopium
(mountain whitefish).  If Corbicula is removed along with the unionids
as EPA proposed (and which I disagree with), then Deltistes (Lost River
sucker) becomes the 4th most sensitive genera.  Either way,
invertebrates no long occupy the four most sensitive genera.  Although I
did not do a comprehensive check, it appears that the other acute values
in the draft document are consistent with the 1999 document.

Temperature Equation for Invertebrate and Acute Data should be
re-examined

The temperature-ammonia sensitivity relationship applied by EPA to
invertebrate acute data in the draft document is based on the slope
provided in the 1999 document, based on Arthur et al. (1987), but not
used for the acute criterion in that document.  As the 1999 document
makes clear, the temperature relationships from Arthur et al. (1987)
were based on seasonal tests of experimental stream water, including
condition, life stage, etc. of test organisms collected during each
season.  As Arthur et al. (1987) notes, several water quality
characteristics other than temperature varied with season and, even more
importantly perhaps, organism condition was unlikely to be similar
across seasons and at the least was unknown.  EPA provided what I
believe was a poor rationale for using Arthur’s et al. data for
chronic invertebrate data.  The fact that Arthur did not observe a clear
temperature relationship with fish (which are also cold blooded) would
weaken the case that temperature was what was driving the apparent
ammonia sensitivity in invertebrates in his study (and which Arthur
himself acknowledges in his paper).

The temperature-toxicity relationship derived by EPA for ammonia in the
1999 document had a fairly steep slope; e.g., a difference between 20
and 25( C resulted in 2-3 fold decrease in the LC50.  There are a few
invertebrate species in the draft document for which it appears possible
to reanalyze a temperature toxicity relationship, and to determine if
one actually exists.

I have not had time to do a formal analysis of the invertebrate acute
data, however, my initial reaction is that temperature may be somewhat
weakly related to acute ammonia toxicity in invertebrates.  For example,
for the snail species Potamopyrgus antipodarum, Hickey and Vickers
(1994) reported an LC50 of 8.71 mg N/L at pH = 8.2 and 15( C and an LC50
of 4.08 mg N/L at pH = 8.2 and 25( C (a factor of 2 over a 10( C
difference).  The LC50 they reported at the same pH and 20( C was nearly
identical to the LC50 at 25( C (4.73 mg N/L).  LC50’s for the
fatmucket mussel at pH = 8.1 and 20( C ranged from 5.2 mg N/L to 11 mg
N/L (Wang et al. 2008) a factor of about 2, which could be within the
range specified by a 5( C difference in EPA’s temperature equation. 
Even larger differences in LC50’s are commonly reported for other
invertebrates based on tests at the same pH and temperature (e.g.,
Villosa iris LC50 = 3.0 – 8.9 mg N/L at pH = 8.3 and 20( C).  These
data suggest that a temperature adjustment factor may be premature for
acute invertebrate data.  Acute data for Ceriodaphnia dubia may perhaps
provide the best information for invertebrates thus far.  At pH = 8.2
and temperature of 7( C, the LC50 = 16.65 mg N/L (Nimmo et al. 1989). 
At pH = 8.16 and 22( C, the LC50 = 30.08 mg N/L (Black 2001), however,
at pH = 8.0 and 25( C, the LC50 = 14.52 mg N/L (Scheller 1997), similar
to the LC50 reported at 7( C.  EPA should reanalyze current acute data,
similar to the way they did for the 1999 document, to determine the
relationship between temperature and acute ammonia toxicity to
invertebrates.

Hyalella azteca position statement and proposed rationale (see Appendix
B):

The EPA workgroup developed a position statement and proposed supporting
rationale describing the concerns over using Hyalella azteca toxicity
test data in criteria development due to the uncertain health of the
test organisms in different test water composition.  The rationale
defines the specific concerns and uncertainties supporting the
recommended exclusion of the Hyalella data from use in criteria
derivation, at this time; the position statement is based on the
workgroup’s review of a number of toxicity tests on Hyalella,
referenced in the rationale.  

5.	Are the position statement and supporting rationale regarding use of
toxicity data for Hyalella azteca in criteria development reasonable and
scientifically sound recommendations?  

Response:  I have reviewed the Hyalella data in a previous outside peer
review for EPA and agreed with EPA’s concerns regarding Hyalella test
data for ammonia.  I agree with the position statement in Appendix B
that Hyalella toxicity data should not be used in water quality criteria
development.  There is a clear need to resolve water quality
requirements and other methodological issues for this species in
water-only tests.



  SEQ CHAPTER \h \r 1 Chronic Freshwater Criterion:

6.	Are the toxicity tests and other studies used to derive the criterion
scientifically defensible for such use?  Are you aware of other relevant
data that were not used?

Response:  Most of the chronic toxicity tests relied on by EPA for
chronic freshwater criterion development are consistent with EPA’s
Guidelines and other water quality criteria developed.  I am not
convinced that the 28 day juvenile test data for the snails included in
Table C should be used.  Both of these tests used mixed-aged organisms,
and in the case of the Ozark spring snail, organisms were also
field-collected adding uncertainty in terms of their condition and
acclimation to laboratory test conditions.  In general, EPA has not
relied on such tests for chronic criteria development, and I believe for
good reason.  I am not aware of any other relevant data that EPA did not
already consider in this document.

7.	Is the freshwater chronic criterion scientifically defensible with
mussels present and mussels absent?  

Response:  I don’t think it is.  See answer to question 2.  The
chronic value used for L. fasciola in Table 4 (0.39 mg N/L), and which
must have been used in the GMCV calculation in Table C, doesn’t appear
to make sense given an IC25 at the same pH and temperature reported of
0.38 mg N/L on p. 20 and the IC25 based on the actual data (pH = 8.2 and
20 ( C) of 0.39 mg N/L.  The IC or EC20 should be lower than the IC or
EC25.  The IC pin value of 0.23 mg N/L appears more in line with the
data for this species.  EPA should at least defend why 0.39 mg N/L is a
more appropriate value than 0.23 mg N/L for this species.  In addition,
wouldn’t it be more consistent to use the ICp value for both Lampsilis
species because the L. siliquoidea chronic value could only be
calculated using the ICpin method?

Use of 28-day Juvenile Test Data (see Appendix C):

Water quality criteria for the protection of aquatic life are derived
using toxicity endpoints that relate to population level impacts.  In
general, these endpoints relate to survival, growth and/or reproduction.
 The 28-day test with juvenile mussels, while similar in duration to a
standard chronic test, is not technically an early-life stage test
according to the 1985 Guidelines for Aquatic Life Criteria as much of
the early development will have already occurred.

8.	Given that the juvenile life stage of freshwater mussels is
relatively long (2-6 years) are 28-day exposure tests with juvenile
mussels scientifically defensible as "chronic" test data for criteria
development?   

Response:  The question posed here by EPA implies that a 28-d test may
not be sufficiently long for freshwater mussels that live 2-6 years
(actually, many unionid species live far longer than that—20 years is
not uncommon).  If this is the case, I fail to see how EPA justifies use
of 28-d tests for fish such as salmonids that live 3-5 years or sturgeon
that live > 30 years.  I realize there’s McKim’s 1977 paper but that
was based on outdated test methods and has questionable relevance to
data collected in more recent years.

The discussion concerning the usability of 28-d mollusk survival and
growth tests in chronic criteria development (pages 15-16 of the draft
document) appears to have contradicting statements and questionable
rationale in my view.  The discussion begins with the statement that
28-d tests do not qualify as life-cycle, partial life cycle, or early
life stage tests as defined in EPA’s Guidelines.  According to those
guidelines, 28-d tests can only be used for fish because it has not been
demonstrated that a 28-d exposure is a reasonable predictor of
invertebrate sublethal chronic effects (defined as at least 90d for
vertebrates and presumably for invertebrates such as mussels which live
for several years).  However, the draft document goes on to say that
28-d survival information could be used to inform the chronic criterion
because obviously, lethality is not reversible.  Yet in footnote #2 on
page 15 of the document, as well as in Appendix C, uncertainties
regarding the 28-d mussel test method are discussed which include
optimal quality and quantity of food, exposure apparatus, etc. needed to
sustain juvenile mussels in good health and maintain adequate growth
throughout the test.  Based on my limited experience testing bivalves, I
believe it may not be as challenging as one might expect to keep control
juvenile mussels alive for 28-d given adequate oxygen and clean water
but it may be quite another thing to document that organism condition
(or alternatively, susceptibility to stress) is truly adequate. 
Therefore, the fact that juvenile mussel survival is lower with exposure
to certain levels of ammonia may be due to having an already stressed
population due to inadequate food or other factors.

A similar finding is now being made by EPA after reviewing chronic
Hyalella chronic test data (reported to be one of the most sensitive
species in the 1999 ammonia criteria document).  In that case, control
survival was apparently satisfactory but test organisms were in fact
ultrasensitive to the chemical due to suboptimal test conditions.

A second point I would raise regarding the chronic mussel tests is that
they were only conducted once (based on published information) for a
given species in one laboratory.  While this in itself is not a reason
to exclude test data, it is clear that there is considerable variability
in acute endpoints among juvenile mussel tests using the same species
and conditions (even within the same lab as mentioned previously), which
should apply to chronic survival with ammonia as well.  Given that
certain mollusks appear to be far more sensitive to ammonia than any
species tested thus far, there should be some replication of these
results to be sure we’re not going to find out that EPA needs to do
yet another reassessment because of issues similar to the ones they are
dealing with now for Hyalella.

It is important to note that while interlaboratory testing was conducted
for acute juvenile tests (Wang et al. 2007), similar interlab testing
has not been conducted (or at least published) to my knowledge for 28-d
mussel chronic tests.  In addition, one should view the acute
interlaboratory results cautiously because, as mentioned previously, a
single culture and water source was used by all laboratories in that
study.  As Wang et al. (2007) point out:  “However, the present study
used the same batch of test organisms and the same dilution water, which
were different from the referenced inter-laboratory studies and might
have contributed to the lower variability in test results. The present
study was designed to determine the inherent variability in the test.
Higher variability would be likely if the inter-laboratory tests were
conducted with different sources of dilution water and organisms (e.g.,
from different populations or watersheds). Therefore, additional study
is needed to further characterize potential variability associated with
the newly developed ASTM standard methods for conducting acute toxicity
tests with early life stages of freshwater mussels.”  Therefore, the
CVs reported by Wang et al. (2007) are likely to be underestimates of
true test method variability.  This is borne out by the acute data
generated for the same species by different labs as noted previously in
my comments.

9.	Should toxicity studies on the growth rates of mussel shells during
28-day tests be considered quantitatively when developing water quality
criteria?

Response:  I don’t believe there is sufficient information or adequate
test method development and standardization at this time to be able to
use growth rates of mussel shells in 28-d tests for criteria
development.  Growth rates, based on shell measurements, are not
sufficiently large over a 28-d period to have much confidence in the
results.  Based on the studies reported in this draft, as well as others
of which I’m aware, shell length is too coarse a measurement to be
used as an indicator of mussel growth over such a short time period. 
Furthermore, it is generally not known what the expected growth rate
should be for a given species under these test conditions.  As I
indicated earlier, without having such information, I don’t believe
that survival data from such tests can be used with confidence as well.

10.	 Regarding the position statement and rationale on use of juvenile
mussel growth data – 

Is it scientifically defensible to include the juvenile growth data from
a 28-day exposure period as “other data” that might influence the
criteria however not be used directly in the derivation of the criteria
value? 

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

Response:  I agree with EPA’s position that mussel growth data should
be categorized as “other data” until it is demonstrated that there
is a repeatable, defensible methodology for generating adequate growth
data and that the growth endpoint calculated is ecologically meaningful.
 For this reason, I don’t think it matters whether impaired growth
observed in a 28-d test is indicative of effects at longer exposures or
not.  However, I would recommend that EPA not consider 28-d mussel
growth as a valid endpoint based on methodological and perhaps
biological uncertainties, as I mentioned under (9) above, rather than
because reduced growth rate is theoretically a reversible effect.  I see
no reason why reduced growth would not be an even bigger effect with
increased exposure duration to a toxicant.

11.	The values of the acute and chronic ammonia criteria have a strong
dependence on pH.  Juvenile and adult mussels, as sediment-dwelling
organisms, inhabit a medium that may have vertical pH gradients, thereby
creating some uncertainty about the appropriate pH to assign as their
exposure conditions.  For applying a criterion protecting mussels, do
you have suggestions on how states and EPA might determine the pH
applicable to the sediment micro-environment to which mussels are
typically exposed?

Response:  I am not so sure that applying ammonia criteria to juvenile
mussels at least, and possibly adults as well, should consider sediment
pH at all for several reasons.  First, larger juvenile and adult mussels
are reported to be exposed to waterborne contaminants via siphoning the
overlying water and not as much via sediment ingestion or pore water
exposure during their growth season (Strayer et al. 2005).  This is
particularly true during times when water temperature exceeds 50( F
(Watters et al. 2001; Schwal and Pusch 2007).  In fact, many mussel
sampling protocols specify a minimum water temperature of 50-60( F
because mussels are near or on the surface under those conditions and
more readily seen (e.g., Pennsylvania DEP and Army Corps sampling
protocol in the Allegheny River, PA).  During this time (typically
spring, summer, and fall in most waterbodies of the temperate U.S.),
mussels are actively feeding via siphoning the water column and sediment
pH is probably not a factor affecting water column ammonia toxicity. 
Rather it is the overlying water pH that is probably more important.

Second, many researchers report that sediment pH and ammonia can vary
substantially from one location to another within even small areas and
may be dependent on very complex redox gradients that are difficult to
measure much less predict (Burton 1992; Stemmer et al. 1990; Sarda and
Burton 1995).  It is no easy matter to identify an appropriate sediment
pH that should be used to apply an ambient water ammonia criterion.

Third, many of the mussel species tested thus far, or that are species
of concern, inhabit well-aerated sediments that often have either fairly
coarse particle size (gravels – cobbles) or sand, both of which have
interstitial water closely resembling the water quality conditions of
the overlying water (Salmon and Green 1983; Neves and Widlak 1988; Way
et al. 1990).  In these cases one would not expect a large difference
between sediment and overlying water pH.

Fourth, it seems premature to consider complex pH adjustments based on
sediments when it hasn’t been demonstrated that juvenile or adult
mussels are as sensitive to sediment ammonia as they appear to be to
water column ammonia in lab tests.  If indeed sediment pH needs to be
considered, then it seems likely that mussels are not being exposed to
surface water (including surface water ammonia) but rather sediment and
interstitial ammonia.  If this is the case, there is no need to adjust
water column ammonia criteria based on sediment pH.

Finally, basing a water column criterion on sediment water
characteristics represents a major departure for EPA in terms of
criteria development.  While I would encourage the Agency to consider
more holistic criteria that take into account multiple media (e.g.,
sediment and water), I don’t believe there is sufficient information
at this time to warrant altering a water quality criterion, which was
derived using water exposure test methods exclusively, due to sediment
characteristics.

12.	In general, should the criteria include a consideration for the
potential pH difference between sediment and the water?  If so, what is
the most scientifically defensible way to account for these differences
when deriving protective water quality criteria?

Response:  I don’t think criteria should consider sediment pH or the
difference between sediment and water pH in deriving water quality
criteria.  Rather, EPA should establish sediment ammonia criteria, using
toxicity test data for mussels and other benthic species (e.g.,
Hyalella, Hexagenia, etc.) exposed to sediments with ammonia.

13.	Should exposure tests on juvenile mussels be conducted with or
without sediment in the test chamber?  

Response:  Tests using juvenile mussels should be conducted using
sediment.  Several researchers reported better survival and growth of
juvenile mussels with the addition of sediment (e.g., Gatenby et al.
1996, Yeager et al. 1994) and I would note that many of the mussel
culturing facilities that produce juveniles for transplants, use
sediment in their culturing systems.  The ASTM standard for glochidia
and juvenile mussel testing (E 24550-06) indicates this in several
places including sections 10.4.2.4, 10.5.2.1, 10.5.2.5, 10.6.3.3, and
10.6.3.13.  Laboratory studies by Yeager et al. (1994) observed that
young juveniles burrow immediately to < 1cm  deep in the sediment, and
that they were not exposed to overlying water.  Others also have
reported rapid burrowing of new juveniles in sediment.  This information
suggests to me that some sediment in test containers may help ensure
that juvenile mussels are unstressed and respond appropriately to
contaminants.  This might be a good check to see whether ammonia
concentrations shown to be toxic in water-only exposures are also toxic
when the tests are conducted with some sediment (clean, sterile, etc.)
as well.

LITERATURE CITED

Arthur, J.W., C.W.West, K.N. Allen and S.F. Hedtke. 1987. Seasonal
toxicity of ammonia to five fish and nine invertebrates species. Bull.
Environ. Contam. Toxicol. 38(2): 324-331.

Black, M. 2001.Water quality standards for North Carolina's endangered
mussels. Department of Environmental Health Science, Athens, GA.

Burton, G.A.  1992.  Sediment collection and processing factors
affecting realism.  In:  G.A. Burton, Ed., Sediment Toxicity Testing. 
Lewis, Boca Raton, FL, pp. 37-64.

Catherine M. Gatenby, Richard J. Neves and Bruce C. Parker 1996. 
Influence of Sediment and Algal Food on Cultured Juvenile Freshwater
Mussels.  Journal of the North American Benthological Society, Vol. 15: 
597-609

Dodd, BJ, MC Barnhart, CL Rogers-Lowery, TB Fobian, RV Dimock, Jr, 2006.
Persistence of host response against glochidia larvae in Micropterus
salmoides. Fish & Shellfish Immunology 21:473-484

Hickey, C.W. and M.L. Vickers. 1994. Toxicity of ammonia to nine native
New Zealand freshwater invertebrate species. Arch. Environ. Contam.
Toxicol. 26(3): 292-298.

Kernaghan, N. J., Gross, T. S., Bishop, C.D., Wang, N., Roberts, A., and
Ingersoll, C. G.. 2005. “Laboratory Toxicity Testing with Freshwater
Unionid Mussels,” Freshwater Bivalve Ecotoxicology, van Hassel, J. H.,
and Farris, J. L., eds., SETAC Press, Pensacola, FL.

M. M. Yeager, D. S. Cherry and R. J. Neves. 1994. Feeding and Burrowing
Behaviors of Juvenile Rainbow Mussels, Villosa iris (Bivalvia:Unionidae)
Journal of the North American Benthological Society, Vol. 13: 217-222

Neves, R. J., and Widlak, J. C. 1988. Occurrence of Glochidia in Stream
Drift on Fishes of the Upper North Fork Holston River, Virginia.  Am
Midland Naturalist, 119, 1988, pp. 111-120.

Nimmo, D.W.R., D. Link, L.P. Parrish, G.J. Rodriguez, W. Wuerthele and
P.H. Davies. 1989. Comparison of on-site and laboratory toxicity tests:
Derivation of site-specific criteria for un-ionized ammonia in a
Colorado transitional stream. Environ. Toxicol. Chem. 8(12): 1177-1189.

Rogers-Lowery, CL and RV Dimock, Jr.  2006. Encapsulation of attached
ectoparasitic larvae of freshwater mussels by epithelial tissue on fins
of naive and resistant host fish. Biol. Bull. 210:51-63 

Salmon, A. and R.H. Green.  1983.  Environmental determinants of unionid
clam distribution in the Middle Thames River, Ontario.  Canadian Journal
of Zoology 61: 832-838.

Sardia, N. and G. Burton.  1995.  Ammonia variation in sediments: 
Spatial, temporal and method-related effects.  Environ. Toxicol. Chem.
14(9):1499-1506.

Scheller, J.L. 1997. The effect of dieoffs of Asian clams (Corbicula
fluminea) on native freshwater mussels (unionidae). Virginia Polytechnic
Institute and State University, Blacksburg, VA.

Schwal, A. and M. Pusch  2007. Horizontal and vertical movements of
unionid mussels in a lowland river Journal of the North American
Benthological Society 26(2):261–272.

Stemmer, B.L., G.A. Burton, Jr. and G. Sasson-Brickson.  1990.  Effect
of sediment spatial variance and collection method on cladoceran
toxicity and indigenous microbial activity determinations.  Environ.
Toxicol. Chem. 9:1035-1044.

Strayer, D. L., Downing, J. A., Haag,W. R., King, T. L., Layzer, J. B.,
Newton, T. J., and Nichols, S. J. 2004. Changing Perspectives on Pearly
Mussels, North America’s Most Imperiled Animals. BioScience, 54:2004,
pp. 429-439.

U.S. Environmental Protection Agency.  1994.  Interim Guidance on
Determination and Use of Water Effect Ratios for Metals, Appendix B: The
Recalculation Procedure.  EPA-823-B-94-001.  Office of Water,
Washington, DC.

USEPA. 2000.  Ambient Aquatic Life Water Quality Criteria for Dissolved
Oxygen (Saltwater):  Cape Cod to Cape Hatteras.  EPA-822-R-00-012 Office
of Water, Washington, D.C. November 2000 

USEPA. 2002.  Clinch and Powell Valley Watershed Ecological Risk
Assessment. 2002.  EPA-600-R-01-050.  NCEA, ORD, Washington, D.C.

Wang, N., C.G. Ingersoll , D.K. Hardesty, C.D. Ivey, J.L. Kunz, T.W.
May, F.J. Dwyer, A.D. Roberts, T. Augspurger, C.M. Kane, R.J. Neves and
M.C. Barnhart. 2007. Contaminant sensitivity of freshwater mussels:
Acute toxicity of copper, ammonia, and chlorine to glochidia and
juveniles of freshwater mussels (Unionidae). Environ. Toxicol. Chem.
26(10):2036-2047.

Wang, N., J.T. Augspurger, M.C. Barnhart, J.R. Bidwell, WG. Cope, F.J.
Dwyer, S. Geis, I.E. Greer, C.G. Ingersoll, C.M., Kane, T.W. May, R.J.
Neves, T.J. Newton, A.D. Roberts and D.W. Whites. 2007. Contaminant
sensitivity of freshwater mussels: Intra- and inter-laboratory
variability in acute toxicity tests with glochidia and juveniles of
freshwater mussels (Unionidae). Environ. Toxicol. Chem.
26(10):2029-2035.

Watters, G. A., 2005. A Brief Look at Freshwater Mussel (Unionacea)
Biology. Freshwater Bivalve Ecotoxicology, van Hassel, J. H., and
Farris, J. L., eds., SETAC Press, Pensacola, FL.

Watters, G. T., S. H. O’dee, And S. Chordas. 2001. Patterns of
vertical migration in freshwater mussels (Bivalvia: Unionidae). Journal
of Freshwater Ecology 16:541–549.

Way, C.M., A.C. Miller and B.S. Payne.  1990.  The influence of physical
factors on the distribution and abundance of freshwater mussels
(Bivalvia:Unionidae) in the lower Tennessee River.  The Nautilus 103(3):
96-98.



PEER REVIEW COMMENTS FROM

George Dixon, Ph.D.

Professor, Department of Biology

and

Vice President of University Research

University of Waterloo

Waterloo, ON  Canada



11 August 2009

Peer Review of “Draft Reassessment of the 1999 Ambient Water Quality
Criteria for Ammonia – Freshwater” by D. George Dixon. EPA Contract
No. EP-C-07-059. Work Assignment No. 1-04.

I have reviewed the document listed above as requested in the Charge to
Reviewers dated 24 July 2009. In addition I have gone through the
background document “1999 Update of Ambient Water Quality Criteria for
Ammonia”. Let me state at the outset that I find the Draft
Reassessment to be a well conceived, well developed and carefully
written document. The acute and chronic ammonia criteria for freshwaters
with mussels present, and separate criteria for waters with mussels
absent, are well founded and scientifically defensible. I have presented
my review comments as responses to the thirteen questions posed in the
Charge to Reviewers Those questions are repeated here for clarity. 

Acute criteria in fresh waters:  

1. Are the toxicity tests used to derive the criteria scientifically
defensible for such use?  Are you aware of other relevant data that were
not used?

I am familiar with the majority of the literature cited and, in my
opinion, the tests used to develop the acute criteria are scientifically
defensible. I am, however, relying on the fact that the Draft
Reassessment states that a process to identify data meeting the minimum
requirements for inclusion was followed.  I have not obtained all of the
cited literature and reviewed each paper for suitability of inclusion of
the findings in the data sets; this type of detailed audit is beyond the
scope of this review. I follow the ammonia literature closely, and I am
unaware of any additional data that could be of use.

2.  What are the technical considerations that EPA should evaluate when
mussels are present and mussels are absent with respect to the
recommended acute criteria?

The document is silent on the appropriate methods for determining
whether mussels are present or absent in a given receiving environment,
and therefore which criteria should be applied. I expect that a guidance
document will have to be developed to assist those applying the revised
criteria. My expertise is in aquatic toxicology, while the question of
presence or absence is based more in field-based aquatic ecology,
particularly mussel ecology.    As such I do not have detailed
recommendations for the determination. I would, however, expect that
some consideration would be given to a method that was able to deal with
environments where mussels were once present, but have now been
expatriated. 



3.	Is it scientifically defensible to exclude the glochidia data at this
time due to the uncertainty of appropriate test duration time for this
life stage?  Do you believe there is an alternative approach to the use
of this data that would be more scientifically sound?

Exclusion of the glochidia data, as detailed in Appendix A, is fully
appropriate and well explained.  Until tests are completed to determine
(on a species by species basis) the duration of the period of longevity
(and by inference viability) for glochidia, LC50s obtained with those
life stages will be suspect. I am unaware of any other approach which
would be more scientifically valid 

 

4.	Regarding the proposed approach to glochidia data in the 2009 draft
position statement as it relates to ecological relevance and
practicality - Is the approach a scientifically defensible principle for
structuring the population exposure duration problem and designing
further research to quantify such a duration?

I am not a mussel ecologist, but the approach taken, and its
implications for further research appear valid. Having said that, it is
likely that modifications to the approach will have to be made as such
research proceeds. It is also apparent that there is sufficient species
variability in the life habits of glochidia that a single approach will
not fit all species. 

Hyalella azteca position statement and proposed rationale (see Appendix
B):

The EPA workgroup developed a position statement and proposed supporting
rationale describing the concerns over using Hyalella azteca toxicity
test data in criteria development due to the uncertain health of the
test organisms in different test water composition.  The rationale
defines the specific concerns and uncertainties supporting the
recommended exclusion of the Hyalella data from use in criteria
derivation, at this time; the position statement is based on the
workgroup’s review of a number of toxicity tests on Hyalella,
referenced in the rationale.  

5.	Are the position statement and supporting rationale regarding use of
toxicity data for Hyalella azteca in criteria development reasonable and
scientifically sound recommendations?  

The exclusion of Hyalella azteca data, as summarized on page 13, is
fully defensible. I would suggest that while all of the points are
relevant, the pivotal consideration is the impact of chloride and
bromide ion on the viability of the species and the potential
interaction of the concentration of these ions with ammonia toxicity.
The Hyalella data will be suspect until that information is obtained.

  SEQ CHAPTER \h \r 1 Chronic Freshwater Criterion:	

6.   Are the toxicity tests and other studies used to derive the
criterion scientifically          

      defensible for such use?  Are you aware of other relevant data
that were not used?

Please see the response to question 1 above. That statement is also
applicable to the chronic data base.

7.	Is the freshwater chronic criterion scientifically defensible with
mussels present and mussels absent?  

The chronic criterion is defensible, provided that suitable guidelines
are developed to differentiate environments where mussels are present
from those where they are absent. Please see my response to question 2
above. 

Use of 28-day Juvenile Test Data (see Appendix C):

Water quality criteria for the protection of aquatic life are derived
using toxicity endpoints that relate to population level impacts.  In
general, these endpoints relate to survival, growth and/or reproduction.
 The 28-day test with juvenile mussels, while similar in duration to a
standard chronic test, is not technically an early-life stage test
according to the 1985 Guidelines for Aquatic Life Criteria as much of
the early development will have already occurred.

8.	Given that the juvenile life stage of freshwater mussels is
relatively long (2-6 years) are 28-day exposure tests with juvenile
mussels scientifically defensible as "chronic" test data for criteria
development?   

This is well covered in Appendix C, and the arguments are convincing. I
agree that the 28-d test does not qualify as a chronic test, since it is
neither a  life-cycle test nor a partial life-cycle test. I also agree
that a concentration causing a greater than 20% reduction in survival
could be used as an upper limit for the Species Mean Chronic Value,
assuming the test met the requirement for good lab practice. 

9.	Should toxicity studies on the growth rates of mussel shells during
28-day tests be considered quantitatively when developing water quality
criteria?

The approach taken in the document (Appendix C), that of not using data
for growth of organisms in non-life-cycle tests with durations of less
than 90 days in the derivation of a Species Mean Chronic Value (SMCV),
is in my opinion perfectly appropriate. The justification given in
paragraph 5 of Appendix C centers on two points: lack of direct
comparisons of growth between shorter tests and life-cycle tests and
uncertainties around the relationship between short-term growth
reductions and growth effects over a full life cycle. Both are valid
reasons to exclude the data from calculations of SMCV.

10.	 Regarding the position statement and rationale on use of juvenile
mussel growth data – 

Is it scientifically defensible to include the juvenile growth data from
a 28-day exposure period as “other data” that might influence the
criteria however not be used directly in the derivation of the criteria
value? 

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

With respect to the first point above, inclusion of the data as “other
data” is perfectly appropriate. This is consistent with the practice
for short-term growth results with other species, and I see no reason
why mussel data should be treated differently. With respect to the
second question, this is a statistically valid point, but it does not
reduce our level of uncertainty around the actual meaning of the data.
The issue will only be resolved when someone undertakes 90-day tests. 

11.  The values of the acute and chronic ammonia criteria have a strong
dependence on pH. Juvenile and adult mussels, as sediment-dwelling
organisms, inhabit a medium that may have vertical pH gradients, thereby
creating some uncertainty about the appropriate pH to assign as their
exposure conditions.  For applying a criterion protecting mussels, do
you have suggestions on how states and EPA might determine the pH
applicable to the sediment micro-environment to which mussels are
typically exposed?

While mussels live in sediment, the pH of importance from an ammonia
criteria perspective is the pH of the respiratory environment, which in
this case is the pH at the sediment-water interface. The sediment pH is
of importance only to the extent that it influences the pH of the water
at the interface. Remember these are water quality criteria, not
sediment quality criteria. I have no suggestion as to the appropriate
method for determining the pH at the interface. The methodology, more in
the realm of environmental chemistry, falls outside of my area of
expertise. 

12.	In general, should the criteria include a consideration for the
potential pH difference between sediment and the water?  If so, what is
the most scientifically defensible way to account for these differences
when deriving protective water quality criteria?

My opinion here is partially covered in the response to question 11. The
pH of the sediment is only of consequence to the degree that it
influences the pH in the water column at the sediment-water interface. I
am not aware of any methodologies, other than direct measurement, that
would allow estimation of the interface pH if the only data available
were the sediment pH and the water-column pH.

13.	Should exposure tests on juvenile mussels be conducted with or
without sediment in the test chamber?  

 

My comments here are limited to laboratory tests where issues of pH
variability between the sediment and the water column could be
eliminated (to avoid direct measurement as discussed above) by
experimental manipulation. The main reason for including sediment would
presumably be to reduce extraneous stress on the experimental organisms,
stress which could possibly interact with the toxicant to produce an
anomalous result. I have no opinion as to whether the presence or
absence of sediment would stress juvenile mussels in an experimental
setting. This will have to be determined experimentally. 



PEER REVIEW COMMENTS FROM

  

Jerry L. Farris, Ph.D.

Professor of Environmental Biology

Arkansas State University

Jonesboro, AR 

EPA’s Draft Reassessment of the 1999 Ambient Water Quality 

Criteria for Ammonia – Freshwater

Acute criteria in fresh waters:

1.	Are the toxicity tests used to derive the criteria scientifically
defensible for such use?  Are you aware of other relevant data that were
not used?

With greater recent attention dedicated to standardization of test
techniques suitably addressing the feasibility of repeatability and
precision of results from juvenile mussel responses (Augspurger, 2007),
data available at the time of this reassessment meet a higher standard
of acceptability for guidance modified from the US EPA (Stephan et al.,
1985) and follow consensus test protocols and quality assurance
recommendations sufficient to support water quality criteria
development.   The majority of tests used to derive the criteria seem
scientifically defensible for such use.  Upon review of those included
and others not specifically included in the calculation, but cited in
Table 1 of the draft addendum, there are tests that present concern for
acclimation procedures of test organisms and culturing (either reported
or unreported) and lack of reference testing that can insure level of
expertise or consistency associated with such field collection, organism
transfer, and culturing that can significantly influence test responses.
 Specifically, sufficient acclimation information for water quality and
temperature was not included in Mummert et al., 2003 and Scheller, 1997.
 

I am not aware of other data from standardized toxicity tests with
ammonia that could be considered relevant to this reassessment.  

What are the technical considerations that EPA should evaluate when
mussels are present and mussels are absent with respect to the
recommended acute criteria?

With mussels presenting such an extreme sensitivity and complex life
history in comparison to other groups, not only will their presence
establish consideration for site variance demonstrations related to
specific media, habitat or host fish interactions, but their spatial
positions and ecological requirements against the backdrop of a
stream’s ability to meet water quality standards- will require some
unique approaches to:

Short- term authorizations to discharge

Compliance with permits that account for ammonia levels, but in streams
or those segments that fail to meet water quality standards

Designated mixing zones and how “shock effect” associated with
ammonia will be considered 

The applicability of biocriteria and use of in-stream community
assessments related to the concept of mussels present and absent that
would incorporate differences between degree of colonization and
distribution rates for fish, other benthic organisms of conventional use
in stream surveys, and freshwater mussels.   These considerations
coupled with only a recent attempt to describe how certain stream
characteristics may serve to structure mussel communities, will need to
be carefully approached when and if such recommendation is made.  

 

Since many of the ‘mussel present’ instances may involve listed
species, the ability to qualify broad testing and culturing of suitable
surrogate species for the time and seasonal limits related to any
testing related to permit or enforcement actions, will likely present an
additional set of laboratory qualifying or certification considerations
related to accessing suitably sited mussel species for testing. 
Laboratories listed in reference to the range of studies cited with this
reassessment and that have implemented the new test standards (ASTM,
2006) have certainly advanced the methodologies needed for risk
assessments and management decisions.  However, this relatively limited
number of research and agency laboratories will likely be challenged to
fulfill the rigorous requirements of consistency with organism culturing
and quality assurance that may be expected of more conventional test
demands.  While consideration of these outcomes may reach beyond the
scope of the question for mussel presence, increased demand as a
technical consideration would be expected to impact population
availability.  

Relative use of reference sites in determining relationships between
areas of reduced freshwater mussel diversity and abundance and known
significant ammonia sources will also likely become a technical
consideration.  Definitive cause-and-effect relationships have not been
documented in such instances (Bartsch et al., 2003) and such
considerations would require more detailed attention between the field
exposure and effects data.    

3.	Is it scientifically defensible to exclude the glochidia data at this
time due to the uncertainty of appropriate test duration time for this
life stage?  Do you believe there is an alternative approach to the use
of this data that would be more scientifically sound?

Information from tests with glochidia can be supportive of the
consideration for site variance information, but should not be used to
provide stand alone estimates in the absence of supportive acute
juvenile estimates.  Multiple lines of evidence comprised of testing
with more than one mussel life stage that include comparisons with
glochidia responses, would seem more scientifically sound, and may
remove much of the seasonal dependence upon culturing and provision of
test organisms with less invasive techniques in instances mentioned in
the above mentioned technical considerations.    



4.	Regarding the proposed approach to glochidia data in the 2009 draft
position statement as it relates to ecological relevance and
practicality - Is the approach a scientifically defensible principle for
structuring the population exposure duration problem and designing
further research to quantify such a duration?

The support of life history information and exposure information can
support this contention, but the pertinence of this endpoint would
continue to be challenged for its scientific defensibility as relates to
what we currently understand of exposure durations related to this life
stage and variety of adaptations.  

Hyalella azteca position statement and proposed rationale (see Appendix
B):

The EPA workgroup developed a position statement and proposed supporting
rationale describing the concerns over using Hyalella azteca toxicity
test data in criteria development due to the uncertain health of the
test organisms in different test water composition.  The rationale
defines the specific concerns and uncertainties supporting the
recommended exclusion of the Hyalella data from use in criteria
derivation, at this time; the position statement is based on the
workgroup’s review of a number of toxicity tests on Hyalella,
referenced in the rationale.  

5.	Are the position statement and supporting rationale regarding use of
toxicity data for Hyalella azteca in criteria development reasonable and
scientifically sound recommendations?  

Those points are very critical given the consideration for
osmoregulation by this organism and the points cited in the
workgroup’s review.  This seems well illustrated in the background
material furnished for this review.

  SEQ CHAPTER \h \r 1 Chronic Freshwater Criterion:

6.	Are the toxicity tests and other studies used to derive the criterion
scientifically defensible for such use?  Are you aware of other relevant
data that were not used?

The tests utilized for the criterion are scientifically defensible for
such use and I’m not aware of other relevant data that were not used. 
  

7.	Is the freshwater chronic criterion scientifically defensible with
mussels present and mussels absent?  

The criterion does not seem defensible for locations where mussels are
present as would relate to considerations under the Endangered Species
Act, but would likely be defensible under considerations with the Clean
Water Act.   I pose this statement as an example of the difficulty in
relying on scientific defensibility in the face of a very broad range of
considerations in risk and decision management that may rely on these
estimates.  

Use of 28-day Juvenile Test Data (see Appendix C):

Water quality criteria for the protection of aquatic life are derived
using toxicity endpoints that relate to population level impacts.  In
general, these endpoints relate to survival, growth and/or reproduction.
 The 28-day test with juvenile mussels, while similar in duration to a
standard chronic test, is not technically an early-life stage test
according to the 1985 Guidelines for Aquatic Life Criteria as much of
the early development will have already occurred.

8.	Given that the juvenile life stage of freshwater mussels is
relatively long (2-6 years) are 28-day exposure tests with juvenile
mussels scientifically defensible as "chronic" test data for criteria
development?   

Research that validates effective endpoints related to the range of
pollutant effects in freshwater mussels have shown that 28-30 -day
exposure tests in both field and laboratory settings can be included to
predict long term impacts related to biomarker responses,
bioaccumulation, and even comparisons to conventional “chronic” test
data.   The 28-day test exposure period will continue to be temporal as
relates to the life cycle, but can be relational to long-term
considerations and offer critical information when combined with
multiple lines of evidence.  Inclusion of more than one life stage
response also advances mussel toxicity assessments, rather than
providing semantic arguments over longevity for a life stage and whether
the exposure period qualifies as “chronic”.   Data from the 28-day
test can be defensible when supported by “other” data.  Other data
should also be scientifically defensible.    

9.	Should toxicity studies on the growth rates of mussel shells during
28-day tests be considered quantitatively when developing water quality
criteria?

Growth rates of mussels during the 28-day tests should not be considered
quantitatively until more information is available on the validity of
this test.  Variability is still attributable to diet and holding
conditions and has been cited broadly.  

10.	 Regarding the position statement and rationale on use of juvenile
mussel growth data – 

Is it scientifically defensible to include the juvenile growth data from
a 28-day exposure period as “other data” that might influence the
criteria however not be used directly in the derivation of the criteria
value? 

Growth should not be considered as defensible for the above mentioned
reasons relating to lack of reliable test conditions, but can be used as
supportive information that might influence the criteria.  Studies that
include recovery and or stability in growth or even degrowth with
freshwater bivalves, have produced more reliable estimates of pollutant
effects and could therefore be used to influence the criteria when
included with broader scale studies.    

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

As stated, the above consideration seems very confusing and should be
clarified to relate to the defensibility as considered with “other
data”. 

11.	The values of the acute and chronic ammonia criteria have a strong
dependence on pH.  Juvenile and adult mussels, as sediment-dwelling
organisms, inhabit a medium that may have vertical pH gradients, thereby
creating some uncertainty about the appropriate pH to assign as their
exposure conditions.  For applying a criterion protecting mussels, do
you have suggestions on how states and EPA might determine the pH
applicable to the sediment micro-environment to which mussels are
typically exposed?

Even if an adequate amount of information was available for any benthic
invertebrates for chronic toxicity related more specifically to the
generic relationship of ammonia toxicity to pH, larger issues challenge
this consideration with the reassessment for changes related to pH
shifts, which may increase when phase waters are isolated for testing. 
These relationships have only recently been examined for juvenile
freshwater mussels, without sufficient detail furnished as to the pH
adjustment used with diluter systems in those tests or their effect upon
alkalinity changes measured during the assays (Wang et al., 2007). 
Although these tests seem suitably performed to establish the influence
of pH on the acute toxicity of ammonia and even the pertinence of
normalizing ammonia toxicity data for mussels to a common pH end point
(total ammonia nitrogen at pH 8.0), information is still lacking for the
sediment relations specific to mussel responses.   As relates to this
consideration, the bulk of the studies that have examined sediment and
phase water relationships to date rely heavily on marine field
monitoring, and not standardized laboratory testing or field monitoring
with freshwaters (Salazar and Salazar, 2007).  Any scientifically
defensible way to account for these differences would greatly benefit
from studies with comparable methods examining habitat preferences,
feeding, and mobility of a sensitive mussel life stage (juvenile) among
sediment compartments, since the synthesis of ammonia toxicity data
confirmed its status as a sensitive species.  Data from these studies
would also be needed to generate pertinent modeling of the ammonia
fluxes from sediment and the concentration profiles in the phases that
would have consequence to mussel impact.     



12.	In general, should the criteria include a consideration for the
potential pH difference between sediment and the water?  If so, what is
the most scientifically defensible way to account for these differences
when deriving protective water quality criteria?

In addition to the aforementioned comments regarding the need for
additional studies involving mussel responses before including a
consideration for this difference, site-specific variance procedures
inclusive of sediment characteristics for a “mussels present “ site
might be especially applicable in this instance.  Should these site
criteria account for such differences in testing of site waters with and
without sediment, they would be expected to be applicable for the
species (or suitable surrogate) and sediment of interest.  Use of a
site-specific criterion for copper that incorporated such broad test
considerations has been utilized to insure protection of mussel species
for almost two decades within the Clinch River, Virginia (Van Hassel,
2007).     

13.	Should exposure tests on juvenile mussels be conducted with or
without sediment in the test chamber?  

If the need is greater at this point with the reassessment to consider
data from tests that offer repeatability and precision of results, than
to ensure inclusion of test conditions reflecting the life history and
habitat selection for this life stage of the mussel, then the test
should exclude the sediment considerations.  There is a growing body of
evidence from tests without sediment that suggests comparable juvenile
shell growth can be measured in control treatments in static,
static-renewal, and flow-through conditions.  This is somewhat
contradictory to earlier studies that recognized the importance of
sediment in distinguishing effects during longer term exposures with
freshwater mussels and fingernail clams.  For this reason, the
information still seems lacking in determining how much of a confounding
effect is introduced by the effects of sediment characteristics (e.g.,
particle size, microbial mobilization and organic carbon content). More
specific information on the survival and physiological condition of
benthic organisms when tested in a water-only medium seems necessary for
testing beyond 96 hours.  Burton et al. (2002) describes a range of
toxicity assays using benthic invertebrates in sediment-free systems
such as interstitial water, elutriate phase, or spiked waters, with each
possibly differentiating toxicant uptake, pathway, and hazard.  It would
seem prudent to be consistent with current conventional approaches to
sediment evaluations for their consideration of appropriate phases to be
tested (extractable, elutriate, interstitial, whole sediment, and in
situ) relating to appropriate conditions for the species or substrate of
interest.  The complex interactions between mussel, water, and sediment
that influence their responses has been cited to reinforce those
considerations (Thorsen et al., 2007). 

	

References:

American Society for Testing and Materials. 2006. Standard guide for
conducting laboratory toxicity tests with freshwater mussels. E2455-06.
In Annual Book of ASTM Standards, Vol 11.06. Philadelphia, PA.

Augspurger, T. 2007.  Advances and opportunities in assessing
contaminant sensitivity of freshwater mussel (Unionidae) early life
stages.  Environ Toxicol Chem 26:2025-2028.

Bartsch MR, Newton TJ, Allran JW, O’Donnell JA, Richardson WB. 2003.
Effects of pore-water ammonia on in situ survival and growth of juvenile
mussels (Lampsilis cardium) in the St. Croix

riverway, Wisconsin, USA. Environ Toxicol Chem 22:2561–2568.

Burton, GA, PM Chapman, and EP Smith.  2002.  Weight of evidence
approaches for assessing ecosystem impairment.  Human and Ecological
Risk Assessment: An international Journal 8:1657-1673.

 

Cherry, DS, JL Scheller, N.L. Cooper and J.R. Bidwell. 2005. The
potential for Asian clam (Corbicula fluminea) dieoffs to impact native
freshwater mussels (Unionidae) I: Water column ammonia levels and
ammonia toxicity. Journal of the North American Benthological Society,
24:369-380.

Cooper, NL. JR Bidwell and DS Cherry. 2005. The potential for Asian clam
(Corbicula

fluminea) die-offs to impact native freshwater mussels (Unionidae) II:
Pore-water ammonia. Journal of the North American Benthological Society,
24:381-394. 

Mummert, AK, Neves RJ, Newcomb TJ, Cherry DS. 2003. Sensitivity of
juvenile freshwater mussels (Lampsilis fasciola, Villosa iris) to total
and un-ionized ammonia. Environ Toxicol Chem 22:2545–2553.

Newton TJ, Allran JW, O’Donnell JA, Bartsch MR, Richardson WB. 2003.
Effects of ammonia on juvenile unionid mussels (Lampsilis cardium) in
laboratory sediment toxicity tests. Environ Toxicol Chem 22:2554–2560.

Salazar, MH and SM Salazar.  2007.  Linking bioaccumulation and
biological effects to chemicals in water and sediment:  A conceptual
framework for freshwater bivalve ecotoxicology.  In Farris JL, Van
Hassel JH, eds, Freshwater Bivalve Ecotoxicology.  SETAC, Pensacola, FL,
USA, pp 215-250.

Scheller JL. 1997. The effect of die-offs of Asian clams (Corbicula
fluminea) on native freshwater mussels (Unionidae). Masters thesis.
Virginia Polytechnic Institute and State University, Blacksburg, VA,
USA.

Stephan CE, Mount DI, Hansen DJ, Gentile JH, Chapman GA, Brungs WA.
1985. Guidelines for deriving numerical national water quality criteria
for the protection of aquatic organisms and their uses. EPA 822/A85/100.
U.S. Environmental Protection Agency, Office of Research and
Development, Washington, DC.

Thorsen, WA, WG Cope, and D Shea.  2007.  Toxicokinetics of
environmental contaminants in freshwater bivalves.  In Farris JL, Van
Hassel JH, eds, Freshwater Bivalve Ecotoxicology.  SETAC, Pensacola, FL,
USA, pp 169-207. 

Van Hassel, JH.  2007.  Case study: Discrimination of factors affecting
Unionid mussel distribution in the Clinch River, Virginia, USA.  In
Farris JL, Van Hassel J, eds, Freshwater Bivalve Ecotoxicology.  SETAC,
Pensacola, FL, USA, pp 311-331.

Wang N, Ingersoll CG, Hardesty DK, Ivey CD, Kunz JL, Dwyer FJ, Roberts
AD, Augspurger T, Kane CM, Neves RJ, Barnhart MC. 2007. Acute toxicity
of copper, ammonia, and chlorine to glochidia and juveniles of
freshwater mussels (Unionidae). Environ Toxicol 26:2048-2056.

Wang N, Ingersoll CG, Greer IE, Hardesty DK, Ivey CD, Kunz JL, Dwyer FJ,
Roberts AD, Augspurger T, Kane CM, Neves RJ, Barnhart MC. 2007. Chronic
toxicity of copper and ammonia to juvenile freshwater mussels
(Unionidae). Environ Toxicol Chem 26:2048-205.



PEER REVIEW COMMENTS FROM

William L. Goodfellow Jr., MS

EA Engineering, Science, and Technology, Inc.

Sparks, MD 



						

26 August 2009

Ms. Laurie Waite

ERG

110 Hartwell Avenue

Lexington, MA  02421-3136

Re:  Review of EPA’s Draft Reassessment of the 1999 Ambient Water
Quality Criteria for Ammonia-Freshwater.

Dear Ms. Waite:

I have completed my review of the above reference EPA water quality
criteria document for ammonia.  I have found the criteria document to be
well organized and very well thought out.  I also feel that the
scientific justification laid out in the document to be excellent.

I have several questions or comments presented below as well as imbedded
in the Charge Questions section (my comments are presented in Bold).

I do question why only standards consistent with ASTM Standards for
testing (page 14) are the only relevant test procedures that were used
in the document. I believe that the EPA’s Whole Effluent Testing (WET)
toxicity testing guidelines for testing are also relevant. Did this
exclude any additional data? From my review of the references, I do not
believe so.

I also question when the Nile tilapia was place in the other chronic
toxicity data (and I agree this was appropriate), why is the Asiatic
clam included in the data used for development of the criteria (and in
fact it is the third most sensitive species for the acute criterion).
The Asiatic clam is also a nonindigenous species (or exotic). In my
opinion I don’t believe that the Asiatic clam should not be used in
the development of the criteria regardless that this exotic is found
throughout United States. I believe elimination of this species from the
dataset will still generate a protective acute criterion.

The Table A in text of the GMAV for the Mountain Whitefish does not
match Table 1. The Text table reports 12.11 and Table 1 reports 12.09. 
I am expecting that the Table 1 is correct and is what was used for the
FAV. Please verify.

As requested, please find below my addresses as they related to the
Charge Questions as presented in ERG’s Charge to Peer Reviewers:

PEER REVIEW CHARGE QUESTIONS

Acute criteria in fresh waters:

1.	Are the toxicity tests used to derive the criteria scientifically
defensible for such use?  Are you aware of other relevant data that were
not used?  

From my review, I believe that the acute toxicity testing data is
relevant. My only concern is that this document was careful to use data
that was from published references which I believe is a strength of this
document as compared to earlier water quality criteria documents. Since
the only unpublished data used that demonstrates a somewhat low toxicity
response to ammonia was from Keller 1999 for the Pondshell Mussel, maybe
putting a synopsis as an attachment would be warranted.  This would help
make the document transparent as the other Attachments help do.  

The Table A in text of the GMAV for the Mountain Whitefish does not
match Table 1. The Text table reports 12.11 and Table 1 reports 12.09. 
I am expecting that the Table 1 is correct and is what was used for the
FAV. Please verify.

What are the technical considerations that EPA should evaluate when
mussels are present and mussels are absent with respect to the
recommended acute criteria? 

I believe that EPA has adequately defended their position for when
mussels were used or not used for the generation of the two phased acute
criterion.  

I do feel that EPA should include guidance that the lower mussel
criterion should only be used when mussels are currently present in the
water system and not historic or wished to be in the system. I can see
States using the lower mussel numbers without a strong technical reason
to use them because someone wished them to be in the system, etc.  When
States choose to go this way arbitrarily, it will weaken the
acceptability of the criterion for everyone.



3.	Is it scientifically defensible to exclude the glochidia data at this
time due to the uncertainty of appropriate test duration time for this
life stage?  Do you believe there is an alternative approach to the use
of this data that would be more scientifically sound?

Not only is it scientifically defensible, I believe that it is the right
way to go. Too much is

unknown with regards to the biology of glochidia and the associated
toxicology or 

appropriate methods to evaluate the toxicity of various chemicals for
this life stage.

4.	Regarding the proposed approach to glochidia data in the 2009 draft
position statement as it relates to ecological relevance and
practicality - Is the approach a scientifically defensible principle for
structuring the population exposure duration problem and designing
further research to quantify such duration?

See comment for Question 3, the same comment is relevant for Question 4.

Hyalella azteca position statement and proposed rationale (see Appendix
B):

The EPA workgroup developed a position statement and proposed supporting
rationale describing the concerns over using Hyalella azteca toxicity
test data in criteria development due to the uncertain health of the
test organisms in different test water composition.  The rationale
defines the specific concerns and uncertainties supporting the
recommended exclusion of the Hyalella data from use in criteria
derivation, at this time; the position statement is based on the
workgroup’s review of a number of toxicity tests on Hyalella,
referenced in the rationale.  

5.	Are the position statement and supporting rationale regarding use of
toxicity data for Hyalella azteca in criteria development reasonable and
scientifically sound recommendations?  

I believe that the position statement and supporting rationale is
appropriate and justified.

  SEQ CHAPTER \h \r 1 Chronic Freshwater Criterion:

6.	Are the toxicity tests and other studies used to derive the criterion
scientifically defensible for such use?  Are you aware of other relevant
data that were not used?

I believe that they are scientifically defensible for the use in the
development of the chronic 

criterion. I am not aware of other relevant data.



7.	Is the freshwater chronic criterion scientifically defensible with
mussels present and mussels absent?  

Similar to the acute criterion, I believe it is scientifically
defensible to develop the criterion with mussels present and mussels
absent. I do believe that the EPA should have a cautionary note as in
the acute criterion as to when and how to determine if mussels are
present or absent.

Use of 28-day Juvenile Test Data (see Appendix C):

Water quality criteria for the protection of aquatic life are derived
using toxicity endpoints that relate to population level impacts.  In
general, these endpoints relate to survival, growth and/or reproduction.
 The 28-day test with juvenile mussels, while similar in duration to a
standard chronic test, is not technically an early-life stage test
according to the 1985 Guidelines for Aquatic Life Criteria as much of
the early development will have already occurred.

8.	Given that the juvenile life stage of freshwater mussels is
relatively long (2-6 years) are 28-day exposure tests with juvenile
mussels scientifically defensible as "chronic" test data for criteria
development?   

I do not believe that 28 day mussel studies should be considered as
chronic test data for the national database as described in the position
statement and supporting documentation. I believe this is very well
thought out and applaud the developers.

9.	Should toxicity studies on the growth rates of mussel shells during
28-day tests be considered quantitatively when developing water quality
criteria?

The problem with using shell growth rates, it doesn’t necessarily link
to a smaller organism being any less viable or having less reproductive
benefits.  Since mussels are not a commercially or recreational
important organism, I believe that the use of growth rates are not valid
as compared to a fish species or commercially important bivalve.

10.	 Regarding the position statement and rationale on use of juvenile
mussel growth data – 

Is it scientifically defensible to include the juvenile growth data from
a 28-day exposure period as “other data” that might influence the
criteria however not be used directly in the derivation of the criteria
value?

I believe that it should remain in the other data section so the
relevant data can be used in site specific evaluations.

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)? 

I don’t believe that data supports this and such a statement should
not be made. 

11.	The values of the acute and chronic ammonia criteria have a strong
dependence on pH.  Juvenile and adult mussels, as sediment-dwelling
organisms, inhabit a medium that may have vertical pH gradients, thereby
creating some uncertainty about the appropriate pH to assign as their
exposure conditions.  For applying a criterion protecting mussels, do
you have suggestions on how states and EPA might determine the pH
applicable to the sediment micro-environment to which mussels are
typically exposed?

It is much easier to do this for water column organisms. I believe that
the agency will open themselves up for tougher battle if they attempt to
make this issue even more complex than it already is. I believe that
noting that this is the case and stating that pH dependence vs. the
criteria as it relates to mussels was not developed at this time given
the uncertainty about what the appropriate pH assignment should be.  

I also argue whether pH in sediment is a relevant question with regards
to criteria since the management tool for water quality is going to be
primarily the NPDES permit. How will anyone be able to control sediment
pH to manage it?  I am not sure what this does for development of a
defensible criterion.

12.	In general, should the criteria include a consideration for the
potential pH difference between sediment and the water?  If so, what is
the most scientifically defensible way to account for these differences
when deriving protective water quality criteria?

Same comment as for Question 11.

13.	Should exposure tests on juvenile mussels be conducted with or
without sediment in the test chamber?  

I believe this question is very similar to the Hyalella azteca question.
If adequate test results that are representative to the toxicity of
ammonia can be performed in water only, that is how they should be
performed. If sediment is necessary for the test to be performed
adequately, then they should use sediment.  However, it should be
cautioned that when the variable of sediment is added to the toxicity
testing methodology, the complexity of the test increases considerably
as to the interpretation. In my opinion the agency should caution the
use of sediment in the test my lead to positive and negative bias of the
test results. The agency should try to stay with water column only
whenever possible.



I thank you for the opportunity to allow me to be a peer reviewer for
this important freshwater water quality criteria document. If you have
any questions or require additional information, please let me know.

Sincerely,

William L. Goodfellow, Jr.

Vice President and Senior Scientist

Clarification question from EPA for the following comment made by Dr.
Bill Goodfellow:

"I also question when the Nile tilapia was place in the other chronic
toxicity data (and I agree this was appropriate), why is the Asiatic
clam included in the data used for development of the criteria (and in

fact it is the third most sensitive species for the acute criterion).
The Asiatic clam is also a nonindigenous species (or exotic). In my
opinion I don't believe that the Asiatic clam should not be used in the
development of the criteria regardless that this exotic is found
throughout United States. I believe elimination of this species from the
dataset will still generate a protective acute criterion."

EPA Question: “…it sounds like he's questioning why we included the
clam data but then the text I highlighted in red sounds like he's saying
that we should include it - maybe there's a typo and the double negative
sentence is wrong?  If you could get clarification from him for me on
this I would appreciate it.

Clarification from Dr. Bill Goodfellow:

In my opinion the Asiatic clam (an exotic) should not be included in the
database used for development of the criteria.  Rather it should be
placed in the Table with other data.  If the Agency decides to include
the Asiatic Clam in the criteria development, I believe considerable
justification will need to be provided in the written justification of
the criteria document.

Hopefully this makes it clearer. Let me know if you have any other
questions or require additional information.



PEER REVIEW COMMENTS FROM

James J. Pletl, Ph.D.

Chief, Technical Services Division

Hampton Roads Sanitation District

Virginia Beach, VA 



EPA’s Draft Reassessment of the 1999 Ambient Water Quality Criteria
for Ammonia – Freshwater (July 14, 2009)

CHARGE TO PEER REVIEW QUESTIONS

Acute criteria in fresh waters:

1.	Are the toxicity tests used to derive the criteria scientifically
defensible for such use?  Are you aware of other relevant data that were
not used?

	The question requires that the term “scientifically defensible” be
defined within the context of the use or the consequences of the use. 
The use is to support development of water quality criteria that will be
promulgated by states as legally enforceable water quality standards.
Failure to meet water quality standards instream and/or at the edge of
respective mixing zones will result in TMDLs, implementation plans, or
NPDES permit limits for dischargers to the respective stream.  Failure
to meet water quality standards also infers that designated uses are not
being supported.  In the case of ammonia criteria, the designated use is
aquatic life.  The Clean Water Act also states that violations of such
permit limits can result in significant monetary fines and/or jail time
for responsible individuals.  Therefore the use, in this case, is quite
demanding of the data.  

The level of scientific defensibility set by EPA for this use is limited
to that defined by EPA’s 1985 “Guidelines For Deriving Numerical
National Water Quality Criteria For The Protection of Aquatic Organisms
And Their Uses”, hereafter referred to as the Guidelines, and the
published, peer reviewed toxicity test methods used in each case.  The
tests in question appear to meet these standards; however the data
quality objectives set by these documents may not be sufficient for the
data use.  The toxicity test methods and the Guidelines do not require
what are considered to be critical elements of scientifically defensible
toxicity tests.  Acute tests typically only have one quality objective:
minimum survival in controls.  Other than control survival there are no
other quality objectives that must be met for the results of these tests
to be considered reliable.  For example, 1) the methods and Guidelines
do not require a concentration-response curve for data to be deemed
acceptable for any use.  There can be little certainty in a toxicity
test result where response does not change predictably with the
concentration exposure.  If demonstration of this curve is not required
of the test or EPA, data of questionable quality may be used to derive
water quality criteria.  2) Analysis of acute toxicity test data
typically also does not address the intra-treatment variability that
commonly occurs in these tests; providing results with biased estimates
on uncertainty.  3) The Guidelines also do not require replication of
any test result for the most sensitive species.  This would allow a
single test result to significantly bias the calculation of water
quality criteria.  ASTM E2455-06 for testing freshwater mussels
recommends that “different batches of the same species and the same
life stage should be collected and tested over time in order to obtain a
measure of the variability associated with testing the particular
species”.  This goal was not accomplished for the pink mucket, the
third most acutely sensitive genus of the database.  

Aside from universal issues for acute tests 4) there is also concern
that the acute juvenile mussel tests may be biased towards findings of
more sensitivity and toxicity.  The tests proposed to be used for this
reassessment are conducted without sediment in the test vessels.  Newton
and Bartsch (2007) found that without sediment juvenile organisms did
not grow in 96 hour toxicity tests.  This clearly shows that the
organisms are stressed by the lack of sediment in the test vessels. 
This also strongly suggests that the survival endpoints from acute tests
may be biased low (falsely indicating greater toxicity) due to the added
stress of failing to provide sediment in the test vessels.  All of these
issues characterize a significant level of uncertainty which is
indirectly proportional to scientific defensibility.  EPA should address
these uncertainties before using new data to update this criteria
document.

	I am not aware of other relevant data that were not used for this
reassessment.

2.	What are the technical considerations that EPA should evaluate when
mussels are present and mussels are absent with respect to the
recommended acute criteria?

Generally, water quality criteria are designed to protect populations. 
The first step in determining whether mussels are present or absent is
to define a standard for the conditions describing that a population
exists at a site.  This is not a straight-forward exercise and can be
quite difficult as seen in various site-specific criteria projects
around the country.  There must be guidance as to what parameters and
measures of those parameters are to be used to define presence/absence
of a population.  Quantitative evidence of reproduction would be
required at a minimum, as well as the presence of various developmental
stages.  The normal abiotic conditions documented for a species should
also be compared to that of a particular surface water to determine
whether a conclusion of presence or absence is supported.  Certainly the
availability of host species could influence the determination of
absence/presence.  

Once absence/presence is established EPA should only use mussel data to
develop water quality criteria for waters where mussels are deemed
present.  EPA should ensure that conditions of the site where absence
has been concluded have not been anthropogenically changed to a state
where mussels cannot reproduce.  Historic records of water quality and
biological surveys would help support this determination.  It is clear
that EPA and the states will share the burden of the presence/absence
determination and that there will be differences of opinion as to
whether, when absence seems apparent, conditions changed due to human
activity that could have been reasonably precluded or the species simply
never successfully established a population at the site due to
species-specific factors.  

EPA should also determine whether simply the presence of mussels is
important or the presence of specific species is important.  Given that
the sensitivity of mussel species in acute tests spans a six-fold
concentration range the presence of only certain species could have a
similar effect on the resulting water quality criteria; which then could
have significant impacts on water quality management decisions. 
Further, different species have different life cycle durations.  If
individual species were used to develop criteria rather than the
presence/absence of a taxonomic group the differences in life cycles may
also impact criteria implementation.   

3.	Is it scientifically defensible to exclude the glochidia data at this
time due to the uncertainty of appropriate test duration time for this
life stage?  Do you believe there is an alternative approach to the use
of this data that would be more scientifically sound?

	The alternative question would be: Is it scientifically defensible to
include glochidia data at this time due to the uncertainty of
appropriate test duration time for this life stage?  Again, EPA has not
defined standards in the Guidelines for this determination other than
requiring acute tests be at least 48 hours in duration and test
organisms not be fed (except for mysids).  These requirements were not
developed to address the duration of the lifestage unique to glochidia. 
Toxicity data for glochidia appear to show that duration of tests with
this lifestage significantly affect the test results when tests are 24
hours in duration or longer.  Therefore it seems that the glochidia are
stressed in some fashion when tests meet the 48 hour requirement of the
Guidelines.  The test design for water quality criteria development is
intended to measure stress due to the treatment (ammonia) independent of
all other stressors.  Tests of 48 hour duration and more fail to meet
this goal.  Since tests of duration less than 48 hours do not meet the
Guidelines’ requirements but tests of duration equal to or greater
than 48 hours provide biased results it seems that a new standard for
test duration unique to these mussels must be first developed.  However
the literature indicates a great deal of diversity among freshwater
mussels in terms of their glochidia ecology.  The extent of this issue
demands either more detail in this criteria document or more
opportunities for site-specific ammonia criteria implemented by the
States for only the mussel species present.  The States will also
require guidance on how to develop site-specific ammonia criteria based
on the issues unique to glochidia.  

To support the scientific defensibility of EPA’s decision it would be
appropriate to first develop guidance for the appropriate test duration
to be used for developing water quality criteria relative to the
duration of the life stage.  Currently acute test duration commonly is
within an order of magnitude of life stage duration.  To accomplish this
goal for glochidia a test duration of 12 hours or less may be necessary.
 Such a guideline would help defend the scientific defensibility of
using glochidia data for developing water quality criteria.  

One should conclude that EPA’s decision to not use glochidia data is
logical and correct without more information. There is a great deal of
uncertainty associated with test results using glochidia.  The question
that must be answered is whether the level of uncertainty realized by
using glochidia data is so great that it compromises the integrity of
the resulting water quality criteria.  This, in fact, seems the case;
however it should be stressed that scientific defensibility should not
be based on judgment but minimum standards of quality.  Without these
standards of quality the magnitude of uncertainty is unknown and
probability of erroneous actions is elevated.

4.	Regarding the proposed approach to glochidia data in the 2009 draft
position statement as it relates to ecological relevance and
practicality - Is the approach a scientifically defensible principle for
structuring the population exposure duration problem and designing
further research to quantify such a duration?

	As discussed earlier the uncertainties of using juvenile mussel data
for deriving water quality criteria are significant, but EPA has
identified additional concerns associated with tests of glochidia.  The
sensitivity of glochidia as a function of exposure duration, within the
constraints of the Guidelines, precludes the use of glochidia data to
derive water quality criteria.  This is primarily due to the relatively
short duration of instream exposure that can occur with this lifestage
and the sensitivity of glochidia to that duration of exposure.  The
issue of exposure duration instream versus that of lab toxicity tests is
critical to the use of all toxicity data in deriving water quality
criteria and is not unique to freshwater mussels and their glochidia
lifestage.  Decapod crustacean larvae are sensitive to duration of test
exposure without food for durations required by the Guidelines.  The
literature refers to the “Point of No Return” or the PNR (Klaus
Anger, various papers), where larvae will not survive beyond a certain
time period without food even if food is provided to the larvae. 
Toxicity tests of duration approaching the PNR are likely stressing the
test organisms in addition to the toxicant in question.  The same can be
said of glochidia tests with durations of 48 hours and more and likely
durations even greater than 24 hours.

Development of criteria must consider all test factors that might bias
test results in comparison to responses that are expected instream, and
exposure constitutes one of the most important factors.  Exposure is a
function of frequency, magnitude and duration.  The default frequency
assumption for water quality criteria toxicity tests is that exposure is
continuous.  Magnitude of exposure is usually dependant on
concentration, which may also be a function of other factors like pH,
alkalinity, DOC, etc.  The magnitude of exposure for glochidia will also
be affected by the species-specific conditions unique to glochidia of
different species.  For example, some glochidia are released into the
water column singly; some are released in groups encased in a gelatinous
shell where exposure would be mitigated.  Duration of test exposure
should be related to the duration of the lifestage, unless it can be
demonstrated that a different test duration is representative of that
lifestage.  This is a fundamental of the Guidelines although it is not
clearly delineated.  In the case of glochidia 48 hour or longer test
durations have been shown to result in responses not expected instream. 
This is a defensible reason to reconsider use of the acute glochidia
data to derive water quality criteria and defines well the research
needs for this taxonomic group and lifestage.  Shorter duration tests
will be necessary to provide reliable test results for deriving water
quality criteria.  

Hyalella azteca position statement and proposed rationale (see Appendix
B):

5.	Are the position statement and supporting rationale regarding use of
toxicity data for Hyalella azteca in criteria development reasonable and
scientifically sound recommendations?  

	The position statement is written to address the use of H. azteca data
for the ammonia update; however the issues presented would likely be
important for many other water quality criteria using data for this
species.  Therefore the position statement is not reasonable and
scientifically sound if H. azteca data are considered within the context
of all water quality criteria calculations.  There is a concern that
decisions such as this do not have associated rules as normally defined
in the Guidelines.  EPA is being reasonable and responsible in
considering factors that may bias the development of water quality
criteria in this case, but it is unclear whether EPA has a process to
identify such factors as new information is collected over time. 
Perhaps there was insufficient data to identify the H. azteca issues
when data for this species was considered for use, but EPA needs to
develop and implement a system to periodically review data for issues
such as this and publish the system as part of the Guidelines.  

Based on this comment the position statement may be reasonable but the
scientific basis of the recommendation is questionable without a well
defined system of review.  Such a system should include review of data
relative to standards of quantity and quality required to make such an
assessment.  EPA needs to determine how much information is necessary to
make this type of recommendation as well as the qualitative nature of
the data (measured vs nominal concentrations, static vs flow through
tests, etc.)  Without a system it will become increasingly difficult to
be consistent in making these types of decisions as more variances
arise.

  SEQ CHAPTER \h \r 1 Chronic Freshwater Criterion:

6.	Are the toxicity tests and other studies used to derive the criterion
scientifically defensible for such use?  Are you aware of other relevant
data that were not used?

	Using the standards established by the Guidelines and the respective
toxicity test methods the toxicity tests used to derive the chronic
criterion are scientifically defensible.  However, this does not mean
that the data generated from these tests and the resulting test
endpoints are scientifically defensible for the use.  As stated before,
there are standards that should be met to establish that the test
results are of sufficient quality to be used in deriving water quality
criteria.  These standards include an appropriate concentration response
curve and limits on variability within and between test treatments,
replication of tests and use of reference toxicant tests to gauge
organism sensitivity, as well as defensible control response across
tests.  One of the shortcomings of the chronic freshwater mussel method
is that it does not include minimum organism size (weight, length)
requirements that must be met at the end of the test in controls.  It is
unknown whether the growth expressed in the controls of these tests is
acceptable in a natural environment where stress does not exist.  The
tests only can establish weights attainable in controls in the lab
environment when organisms in the controls meet the minimum survival
requirement.  The organisms may be stressed in all test vessels,
resulting in a bias of the test endpoint.  The combination of treatment
stressors (ammonia and vessels without sediment, for example) may
produce a synergistic effect resulting in a lower ammonia IC20 for
growth and survival than that resulting from a test which has only one
stressor (ammonia).  In this case, the control response cannot account
for the synergism, even though the control is intended to account for
all factors independent of the tested treatment.

	I am not aware of any other relevant data that was not used.

7.	Is the freshwater chronic criterion scientifically defensible with
mussels present and mussels absent?  

As stated in question #2, EPA should determine whether simply the
presence of mussels is important or the presence of specific species is
important.  The first step in determining whether mussels are present or
absent is to define a standard for the conditions describing that a
population exists at a site.  This is not a straight-forward exercise
and can be quite difficult as seen in various site-specific criteria
projects around the country.  Guidance must be provided as to what
parameters and measures of those parameters are to be used to define
presence/absence of a population.  Quantitative evidence of reproduction
would be required at a minimum, as well as the presence of various
developmental stages.  The normal abiotic conditions documented for a
species should also be compared to that of a particular surface water to
determine whether a conclusion of presence or absence is supported. 
Certainly the availability of host species could influence the
determination of absence/presence.  

Once absence/presence is established EPA should only use mussel data to
develop water quality criteria for waters where mussels are deemed
present.  EPA should ensure that conditions of the site where absence
has been concluded have not been anthropogenically changed to a state
where mussels cannot reproduce and biological records have documented
the presence of previous populations.  Historic records of water quality
and biological surveys would help support this determination.  It is
clear that EPA and the states will share the burden of the
presence/absence determination and that there will be differences of
opinion as to whether, when absence seems apparent, conditions changed
due to human activity that could have been reasonably precluded or the
species simply never successfully established a population at the site
due to species-specific factors.  Given that the chronic sensitivity of
mussel species drives the CCC approximately 44% lower than when mussel
data is not used the management actions and costs based on criteria
without mussel data could be significantly different than when mussel
data is used to develop the CCC.

The presence/absence of species is usually a question that is asked when
criteria are recalculated or a site-specific criterion is being
developed.  In these cases the issue of removing species from a database
is addressed by determining whether data for other species with similar
taxonomy is available.  The taxonomy and reproductive cycle of the
freshwater mussels is so unique that this will not possible.  Therefore
a decision must be made as to whether to apply a CCC with or without
mussel data.  Given that there is so little chronic data available for
these mussels and that there are outstanding issues that must still be
addressed (use of shell measurements to represent growth, tests without
sediment, irregular concentration response curves, etc.) a CCC with
mussel data will result in water quality standards with uncertain
consequences.  This uncertainty demands that criteria without mussel
data be available for states to use when mussels are not resident
(reproducing population found year to year).  At this point in time a
CCC without mussel data would be more scientifically defensible than one
with mussel data.

Use of 28-day Juvenile Test Data (see Appendix C):

8.	Given that the juvenile life stage of freshwater mussels is
relatively long (2-6 years) are 28-day exposure tests with juvenile
mussels scientifically defensible as "chronic" test data for criteria
development?   

	This question is not only a function of the duration of exposure in the
test but whether the test meets the requirements of the Guidelines. 
Clearly the test does not meet the Guidelines requirements because it
does not address either the entire life cycle or, for partial life-cycle
tests, the appropriate life stages and test endpoints defined by the
Guidelines.  The reliability of this test is also clouded by the
apparent impact of testing juveniles in the absence of sediment and the
use of shell length as a chronic test endpoint for freshwater mussels
(see comments below).  Perhaps the Guidelines require updating (has not
been formally revised since its first release in 1985), but until this
is done the 28-day juvenile test does not meet the standard for
scientific defensibility that EPA has established.

	One might suggest using the data from these tests as an estimate of
freshwater mussel chronic toxicity because the tests did not include the
glochidia life stage and it is more appropriate than not using the data
for criteria development.  Unfortunately there is enough concern
regarding the reliability of this test even to accurately represent the
sensitivity of juveniles that it would not be appropriate to use the
data in any water quality criteria capacity.

	Independent of the issues raised above, it would seem that a 28-day
exposure test with juvenile mussels would be representative of the
sensitivity of juveniles.  However, chronic tests used for developing a
CCC are intended to evaluate or extrapolate the sensitivity of a
population to a stressor.  This requires reliable knowledge of the
sensitivity of the most sensitive life stages of that population.  It is
unknown how the sensitivity of glochidia compares to juveniles because
of the issues associated with both acute and chronic test methods. 
Without this knowledge and reasonable certainty it cannot be determined
that the 28-day juvenile test with freshwater mussels is scientifically
defensible relative to development of a CCC.  

9.	Should toxicity studies on the growth rates of mussel shells during
28-day tests be considered quantitatively when developing water quality
criteria?

	Review of Wang et al. (2007) for both the copper and ammonia 28-day
tests shows that there is a very small difference between the length of
the shells at the beginning and at the end of the test based on control
responses.  In fact it appears that some of the tests show no
statistical difference in shell length between the beginning and end of
a test at alpha=0.05 and assuming a normal distribution.  This does not
support the use of shell length to represent growth in the 28-day mussel
tests.  Further this paper shows that using shell length to represent
growth does not provide reliable dose response curves in most cases. 
Although the authors only conducted 21-day tests both Bringolf et al.
(2007a) and Bringolf et al. (2007b) found that the concentration
response curves for different chemicals were not what one would expect
for a toxicant.  Although there may be issues with organisms benefitting
from certain concentrations of chemicals, this fact only complicates the
interpretation of this data and increases the uncertainty of using
toxicity test results based on shell length.

10.	 Regarding the position statement and rationale on use of juvenile
mussel growth data – 

Is it scientifically defensible to include the juvenile growth data from
a 28-day exposure period as “other data” that might influence the
criteria however not be used directly in the derivation of the criteria
value? 

Based on the uncertainty associated with testing juveniles without
sediment and the lack of reliability in the shell length endpoint it is
not appropriate to use the chronic data in any way associated with
development of a CCC. 

Should the statement also consider impaired growth of mussels which were
affected at a 28-day exposure could as likely continue to decline in
longer exposures as another potential outcome (i.e., the chance they
could recover or stabilize is one potential outcome only)?

The comments provided above on the reliability of the shell length
endpoint for chronic tests of freshwater mussels indicate that it is not
a reliable endpoint to represent the chronic sensitivity of these
organisms.  Without more data, it is not defensible to speculate on the
outcome of tests with longer duration when the reliability of the
endpoint being used is in question.

11.	The values of the acute and chronic ammonia criteria have a strong
dependence on pH.  Juvenile and adult mussels, as sediment-dwelling
organisms, inhabit a medium that may have vertical pH gradients, thereby
creating some uncertainty about the appropriate pH to assign as their
exposure conditions.  For applying a criterion protecting mussels, do
you have suggestions on how states and EPA might determine the pH
applicable to the sediment micro-environment to which mussels are
typically exposed?

	It will be necessary for states and EPA to directly measure the pH of
the environment where each life stage exists.  For juvenile and adult
mussels this will require measurements of the sediment in each location
where the criteria are applied as water quality standards.  Guidance
from EPA on how to appropriately represent the pH of sediments where
ammonia criteria are applied to protect freshwater mussels will be
necessary.  Measurements in the field will also be necessary because the
pH of a discharge regulated by a NPDES permit cannot be used to estimate
the pH of the sediment.  The pH of the sediment can only be assessed
through direct measurement.  This may be done in situ or by rapidly
measuring the pH of relevant sediments collected with sediment grab or
core equipment.

12.	In general, should the criteria include a consideration for the
potential pH difference between sediment and the water?  If so, what is
the most scientifically defensible way to account for these differences
when deriving protective water quality criteria?

	The freshwater ammonia criteria must appropriately address the pH of
the juvenile freshwater mussel environment to be reliable.  Although the
criteria must be protective they must also be representative of aquatic
life sensitivities and implemented in a way that does not overstate
those sensitivities.  Being protective is straight-forward using
conservative assumptions but criteria development and implementation
already have multiple layers of conservatism.  These layers include the
return frequency (once in 3 years) of the criteria, the assumption that
exposure is continuous instream for a time period equivalent to the
duration of the tests, an acute criterion averaging period of one hour,
the use of the 90th percentile of temperature and pH to apply the
ammonia criteria, the reasonable potential process of EPA’s Technical
Support Document for Water Quality-based Toxics Control which relies on
a very high percentile (95th or 99th) of the distribution of the highest
concentration measured for an effluent to issue permit limits, the
303(d) listing process which is dependent on the vast minority of data
exceeding the criteria to conclude that a water body is not in
compliance with standards, etc.  The conservatism built into criteria
development and implementation demands that pH instream be accurately
assessed and applied to the criteria to reliably determine compliance
with water quality standards.  

EPA has not adopted data quality objectives to limit the amount of
uncertainty that can exist when applying water quality criteria
instream, therefore a consensus standard to determine if water based
data can be used to estimate sediment based responses is not available
and scientific defensibility cannot be assessed.  However, there is an
approach that will likely provide consensus agreement in scientific
defensibility: to directly address the pH of the instream sediment
within the context of a toxicity test rather than trying to compensate
for pH differences between water tests and the instream sediment.  This
approach will directly address concerns that juvenile test results are a
function of the presence/absence of sediment and avoid the uncertainty
of mechanisms attempting to extrapolate response from water only
exposures.  It is recognized that this recommendation would not allow
the use of current data to derive ammonia water quality criteria
inclusive of freshwater mussel sensitivities.  It would ensure that the
unique ecology of these mussels’ life stages is accurately captured
(within the constraints of conventional testing methods) and that when
the resulting criteria are implemented management actions will be
reliable. 

13.	Should exposure tests on juvenile mussels be conducted with or
without sediment in the test chamber?  

	It is clear from Newton and Bartsch (2007) that the potential for
juvenile freshwater mussels to be stressed in the absence of sediment
exists.  Wang et al. (2007) found that sediment did not appear to impact
shell growth and may have slightly reduced survival, but only one test
using sediment was conducted and this test only consisted of two
replicates per treatment.  As stated previously toxicity tests must
represent the stress experienced by a single stressor to be used in
developing water quality criteria.  Therefore for this particular use,
until more information is available, juvenile tests of any duration must
include sediment in the test vessels in order to be used in deriving
water quality criteria.

LITERATURE CITED

R.B Bringolf, W.G. Cope, S. Mosher, M.C. Barnhart, D. Shea. 2007a. 
Acute and Chronic Toxicity of Glyphosate Compounds to Glochidia and
Juveniles of Lampsilis siliquoidea (Unionidae).  Env. Tox. Chem.26(10),
p.2094

R.B Bringolf, W.G. Cope, M.C. Barnhart, S. Mosher, M.C. Barnhart, P.
Lazaro, D. Shea. 2007b.  Acute and Chronic Toxicity of Pesticide
Formulations (Atrazine, Choropyrifos and Permethrin) to Glochidia and
Juveniles of Lampsilis siliquoidea (Unionidae).  Env. Tox. Chem.26(10),
p.2101

T.J Newton and M.R. Bartsch. 2007.  Lethal and Sublethal Effects of
Ammonia to Juvenile Lampsilis Mussels (Unionidae) in Sediment and Water
Only Exposures.  Env. Tox. Chem. (26(10), p.2057

N. Wang, C.G. Ingersoll, I.E. Greer, D.K. Hardesty, C.D. Ivey, J.L.
Kunz, W.G. Brumbaugh, F.J. Dwyer, A.D. Roberts, T. Augspurger, C.M.
Kane, R.J. Neves, M.C. Barnhart.  2007.  Chronic Toxicity of Copper and
Ammonia to Juvenile Freshwater Mussels (Unionidae).  Env. Tox. Chem.
26(10), p. 2048

 PAGE   6 

  PAGE  ii 

 PAGE   i 

A-  PAGE  2 

A-  PAGE  3 

B-  PAGE  57 

B-  PAGE  56 

C-  PAGE  1 

C-  PAGE  3 

Alex Barron

Alex Barron

C-  PAGE  6 

C-  PAGE  5 

Steven Canton

Steven Canton

Jerome Diamond

Jerome Diamond

George Dixon

George Dixon

Jerry Farris

Jerry Farris

William Goodfellow

William Goodfellow

James Pletl

James Pletl

C-  PAGE  73 

