&EPA
              Umtea Stales
              Environmental Protection
              Agency
                           Office of Water Regulations
                           and Standards
                           Washington. DC 20460
Water Quality
Criteria Methodology

-------
             FINAL REPORT
                  on
WATER QUALITY CRITERIA METHODOLOGY
                   to
U.S. ENVIRONMENTAL PROTECTION AGENCY
       Criteria and Standards Division
             August 6, 1990

         Contract No. 68-03-3534
        Work Assignment No. H2-52
               BATTELLE
  Washington Environmental Program Office
     2101 Wilson Boulevard, Suite 800
         Arlington, Virginia 22201
      Revised October 22,  1990

-------
This report is a work prepared for the United States by
Battelle.  In no event shall either the United States or
Battelle have any responsibility or liability for any
consequences of any use, misuse, inability to use, or
reliance upon the information contained herein, nor does
either warrant  or otherwise represent in any way the
accuracy, adequacy, efficacy, or applicability of the  contents
hereof.

-------
                                1.0 BACKGROUND


     EPA's water quality criteria are developed under the authority of Section

304(a)(l) of the Clean Water Act which reads as follows:


     "The  [EPA] Administrator, after consultation with appropriate Federal and State
     agencies and other interested persons shall develop and publish, within one year
     after  the date of enactment of this title (and from time to time thereafter revise)
     criteria for water quality accurately reflecting  the  latest scientific knowledge (A)
     on the kind and  extent of all identifiable effects on health and welfare including
     but not limited to, plankton, fish,  shellfish, wildlife, plant life,  shorelines,
     beaches, esthetics, and recreation which may be expected from the presence of
     pollutants  in any body of water, including ground water; (B) on the
     concentration and dispersal of pollutants, or their by products, through
     biological,  physical, and chemical processes; and  (C)  on the effects of pollutants
     on biological community diversity,  productivity, and stability, including
     information on the factors affecting rates of eutrophication and rates of organic
     and inorganic sedimentation for varying types of receiving waters."

The Section 304 (a) (1) requirement for development of criteria supports the goal of

fishable swimmable waters given in  Section 101 (a)(2) which states  that:


     "it is  the national goal that, wherever attainable, an interim  goal  of water
     quality which provides for the  protection and  propagation of fish, shellfish, and
     wildlife and provides for recreation in and on the water be achieved by July 1,
     1983."
     The Section 304(a)(l) water quality criteria are issued as guidance to the States.
These criteria are not enforceable regulations.  Rather, these criteria present scientific
data and provide guidance on the environmental effects of pollutants which can be
useful in deriving regulatory requirements based on consideration of water quality
impacts.  Under the Clean Water Act, these regulatory requirements may include the
promulgation of water quality-based effluent limitations under Section 302, water
quality standards under Section 303, or toxic pollutant effluent standards under
Section 307.  The States are responsible for developing State standards.  States may

simply adopt EPA's criteria, modify them on a site-specific basis (i.e., develop site-
specific criteria), or use an entirely different number, provided that the new number is

-------
scientifically defensible.  State standards are submitted to EPA for approval, and if
EPA disapproves, either the State must promulgate a new standard, or in rare cases,
EPA may promulgate standards for the State if the State fails to develop acceptable
standards in a timely manner.  For example, prompted by the  1987 CWA, EPA is
currently promulgating State toxics standards for certain States which have not done
so.
      EPA issues both criteria to protect aquatic life and criteria to protect human
health.  This document confines itself to the methodology for developing  criteria to
protect  human health. The methodology described in this document was published in
the Federal Register on November 28, 1980, "Water Quality Criteria Documents;
Availability, Appendix C - Guidelines and Methodology Used in the Preparation of
Health Effect Assessment Chapters of the Consent Decree Water Criteria Documents"
(45 FR  79347).  This methodology has  been reviewed and  approved by EPA's Science
Advisory Board.  A copy  of these guidelines is included in Appendix A of this
document.
     The purpose  of human  health criteria is to estimate the ambient water
concentration of a pollutant which does not represent a significant risk to the public.
Ambient water quality criteria for human health are primarily based on two types of
biological endpoints: (1)  carcinogenicity and (2) toxicity (i.e., all adverse effects other
than cancer).  Also, criteria may in some cases be based on organoleptic effects
(thresholds for taste or odor).  There are essentially two procedures for assessing
health effects; one which addresses carcinogens and one which addresses  non-
carcinogens.  The reason for having two methodologies is that, for the purpose of
deriving ambient water quality criteria, carcinogenicity is regarded as a non-threshold
phenomenon, whereas toxicity is regarded as having a threshold below which there
will not be an effect.

     Under the assumption that carcinogenicity is a "non-threshold phenomenon,"
there  are no "safe" or "no effect" levels,  because even  extremely small  doses are
assumed to elicit a finite increase in the incidence of the response. Therefore, water

-------
quality criteria for carcinogens are presented as a range of pollutant concentrations
associated with corresponding incremental increases in the risk of developing cancer.
     For compounds which do not manifest any apparent carcinogenic effect, the
assumption used to derive a criterion is that the compound has a threshold below
which no effects will be observed. This assumption is based on the premise that a
physiological reserve capacity (or defense mechanism) exists within the organism
which is thought to be depleted  before clinical disease ensues. Alternatively, it may
be assumed  that the rate of damage will be insignificant over the lifespan of the
organism.  Thus, ambient water  quality criteria are also derived for non-carcinogenic
chemicals, and presumably result in no-observable-adverse-effect levels (NOAELs) in
human populations.
     In some instances, criteria are based on organoleptic characteristics, i.e.,
thresholds for taste or odor.  Such criteria are established when insufficient
information  is available on toxicologic effects, or when the  estimate of the
organoleptic effect level of the pollutant in ambient water is lower than the level
calculated from toxicologic data.  It should be recognized that criteria based solely on
organoleptic effects do not necessarily represent approximations of acceptable risk
levels for human health.  Development of these criteria will not be discussed further
in this document.
     This document describes the methodology used to develop human health criteria
for both carcinogens and non-carcinogens.  The methodology used includes four steps:
hazard identification, dose-response assessment, exposure assessment, and a risk
management decision.  Each of these steps will be  discussed in a separate  section
below.  The criteria developed from this methodology are published in criteria
documents for each of the compounds or groups of compounds evaluated.  These
criteria documents provide the scientific base used  to support development of the
criteria. As discussed in  more detail below, the methodology was not developed to
assess fish contamination directly, but can be utilized for this purpose.  One goal of
this document is to discuss the methodology in detail so that an informed  decision can
be made concerning its utility for this purpose.

-------
                                        4
                          2.0  HAZARD IDENTIFICATION

      Hazard identification involves gathering and evaluating data on the types of
health injury or disease that may be produced by a chemical, and data on the
conditions of exposure under which injury or disease may be produced. It may also
involve characterization of the behavior of a chemical within the body and the
interactions it undergoes with organs, cells or even part of cells.  Data of the latter
types may be of value in answering the ultimate question of whether the forms of
toxicity known to be produced by a substance in one population group or in
experimental settings are also likely to be produced in humans.  Hazard identification
is not risk assessment; rather it is simply the process of determining whether exposure
to an agent can cause an increase in the incidence of an adverse health effect. It
includes an associated characterization of the nature and strength of the evidence of
causation.
      Information for hazard identification can be obtained from EPA's  approved
toxicology data source, the Integrated Risk Information System (IRIS).  IRIS  is an
electronic data base containing health risk and EPA regulatory information on specific
chemicals.   It was developed for EPA staff in response to a growing demand  for
consistent risk information on chemical substances for use in decision-making and
regulatory activities.  The heart of the IRIS  system is its collection of computer files
covering individual chemicals.  IRIS can be accessed to obtain summaries of key
toxicological data to be used in hazard identification.
      The results of the hazard identification process influence the nature and the
extent of subsequent steps in risk  analysis.  For example, the endpoint of concern in a
dose-response assessment may be selected based on the most severe adverse effect
identified in the hazard identification.
      In the hazard assessment,  an attempt is made to include the known relevant
hazard information. The relevant hazard information for a particular compound is
summarized in the  criteria document prepared to support the derivation of the criteria
for that compound.  Review articles and reports are often used in the process of data

-------
                                        5
evaluation and synthesis.  Scientific judgement is exercised in the review and
evaluation of the data and in the identification of the adverse effects for which
protective criteria are developed.  The criteria documents are peer reviewed by a
committee of scientists familiar with the specific compound(s).  These work groups
evaluate the quality of the available data,  the completeness of the data summary, and
the validity of the derived criterion.  As noted below, EPA is not developing any new
human health criteria at the present time.
     In the analysis  and organization of the data an attempt is made to be consistent
with respect to the format and the application of acceptable scientific principles.
Evaluation procedures used in the hazard identification process  follow the principles
outlined  by the National Academy of Sciences in "Drinking Water and Health" (1977),
and the guidelines of the Carcinogen Assessment Group of the U.S.  EPA.
     Chemicals for which human health criteria have been published by EPA include
most of the  126 toxic priority pollutants listed under Section 307 (a) of the Clean
Water  Act (see Table 1).  The method used in developing the priority pollutant list is
unclear; however, toxicity and production  were considered in the selection process.
EPA has  not issued any new human health criteria since 1984,  but will issue new
ones as needed;  existing criteria are being revised to reflect  current science.
     The detection limits  and methods available for analyzing pollutant concentrations
are not considered in setting water  quality criteria.  However, for other purposes, EPA
has spent considerable resources in  developing standard analytical methods for these
pollutants in a number of matrices,  and in extending detection  limits to the state of
the art.  Analytical methods are published in 40  CFR Part 136, "Guidelines
Establishing Test Procedures for the Analysis of Pollutants Under the Clean Water Act"
(49 Federal  Register 2, October 26, 1984). These analytical procedures are used for
compliance monitoring and to express pollutant quantities, characteristics, or
properties in effluent limitations guidelines and ambient water quality standards.

-------
             TABLE 1.  LIST OF 126 PRIORITY POLLUTANTS
   Acenaphthene
   Acrolein
   Acrylonitrile
   Benzene
   Benzidine
   Carbon tetrachloride (tetra-
      chloromethane)
   Chlorobenzene
 * 1,2,4-Tri ch1orobenzene
   Hexachlorobenzene
   1,2-Dichloroethane
   1,1-1-Trichloroethane
   Hexachloroethane
 *1,1-Dichloroethane
   1,1,2-Trichloroethane
   1,1,2,2,-Tetrach1oroethane
 *  Chioroethane
   Bis(2-chloroethyl)ether
   2-Chloroethyl vinyl  ether
      (mixed)
 *  2-Chloronaphthalene
   2,4,6-Tri chlorophenol
 * Parachlorometacresol
   Chloroform(trichloromethane)
   2-Chlorophenol
   1,2-Di ch1orobenzene
   1,3-Dichlorobenzene
   1,4-Dichlorobenzene
   3,3-Dichlorobenzidine
   1,1-Dichloroethylene
   1,2-Trans-dichloroethylene
   2,4-Dichlorophenol
 * 1,2-Dichloropropane
   1,3-Dichloropropylene
 •#  2,4-Dimethylphenol
 *  2,4-Dinitrotoluene
 * 2,6-Dinitrotoluene
   1,2-Diphyeny1hydrazine
   Ethylbenzene
   Fluoranthene
 #  4-Chlorophenyl  phenyl ether
 *  4-Bromophenyl phenyl ether
   Bi s(2-ch1oroisopropyl)ether
«t  Bis (2-chloroethoxy)methane
   Methylene chloride (dichloro-
      methane)
   Vinyl chloride  (chloroethylene)
  Dieldrin
  Chlordane  (tech. mixture
     & metabolites)
  4,4-DDT
  4,4-DDE (p.pDOX)
  4,4-DDE (p.p-TDE)
  Alpha-endosulfan
  Beta-endosulfan
  Endosulfan sulfate
  Endrin
  Endrin aldehyde
  Methyl chloride  (chloro-
     methane)
  Methyl bromide
  Bromoform  (tribromomethane)
  Dichlorobromomethane
  Chlorodibromomethane
  Hexachlorobutadi ene
  Hexach1orocyc1opentadi ene
  Isophorone
4 Naphthalene
  Nitrobenzene
* 2-Nitrophenol
* 4-Nitrophenol
  2,4-Dinitrophenol
  4,6-Di ni tro-o-cresol
  N-nitrosodimethylamine
  N-nitrosodiphenylamine
  N-nitrosodi-n-propylamine
  Pentachlorophenol
  Phenol (4APP method)
  Bis(2-ethylhexyl)phthalate
  Butyl benzyl phthalate
  Di-n-butyl phthalate
* Di-n-octyl phthalate
  Diethyl phthalate
  Dimethyl phthalate
  Benzo(a)anthracene (1,2
     benzanthracene)
  Benzo(a)pyrene(3,4-benzo-
     pyrene)
  3,4-Benzof1uoranthene
  Benzo(k)fluoranthane (11,
     12-benzof1uoranthene)
  Chrysene
  Acenaphthylene
  Anthracene

-------
                               7

           TABLE 1.   LIST OF 126 PRIORITY POLLUTANTS  (Continued)
Aldrin
Benzo(ghi)perylene (1,12-
   benzoperylene)
Fluorene
Phenanthrene
Dibenzo(a.h)anthracene
Indeno(li2,3-cd)pyrene
Pyrene
Tetrach1oroethy1ene
Toluene
Tr i ch1oroethy1ene
Heptachlor
Heptachlor expoxide
Alpha-BHC
Beta-BHC
Gamma-BHC(lindane)
Delta-BHC
PCB-1242
PCB-1254
PCB-1221
PCB-1232
PCB-1248
PCB-1260
PCB-1016
Aroclor 1242
Aroclor 1254
Aroclor 1221
Aroclor 1232)
Aroclor 1248)
Aroclor 1260)
Aroclor 1016)
Toxaphene
Antimony (total)
Arsenic (total)
Asbestos (fibrous)
Beryllium (total)
Cadmium (total)
Chromium (total)
Copper (total)
Cyanide (total)
Lead (total)
Mercury (total)
Nickel (total)
Selenium (total)
Silver (total)
Thallium (total)
Zinc (total)
2,3,7,8-Tetrachlorodibenzo-
   p-dioxin (TCOO)

-------
                                         8
                        3.0  DOSE-RESPONSE ASSESSMENT

     The human health risks of a substance cannot be determined with any degree of
confidence unless dose-response relationships are quantified, even if the substance is
known to be toxic.  Therefore, a dose-response assessment is required before a
criterion can be calculated. The dose-response assessment determines the quantitative
relationship between the amount of exposure to a substance and the onset of toxic
injury or disease.  Data  for determining dose-response relationships are typically
derived from animal studies, or less frequently, from epidemiologic studies in exposed
populations.  Dose-response data for use in calculation of water quality criteria  are
taken from the IRIS system, described in the previous  section.  Data in IRIS are taken
from the literature and peer reviewed  within EPA before entering on the system.
     The dose-response  information needed  for carcinogens is an estimate of the
carcinogenic potency of  the compound. Carcinogenic potency is  defined here as a
general term for a chemical's  human cancer-causing potential.  This term is often used
loosely to refer to the more specific carcinogenic slope factor.  Carcinogenic slope
factor is defined here as an estimate of carcinogenic potency derived using animal
studies or epidemiological  data on human  exposure. It is based on extrapolation from
typical test exposures at high dose  levels over  short periods of time to more realistic
low dose levels over a lifetime exposure period by using a linear model. EPA
generally assumes a 70 year lifetime in these calculations.  The carcinogenic slope
factor is the estimate of carcinogenic potency which is used in developing the criteria.
This estimate of carcinogenic  potency  is generally regarded as a conservative, upper
bound estimate.
     For non-carcinogens, EPA uses the reference dose (RfD)  as  the dose response
parameter in calculating the criteria. The  RfD  was formerly referred to as an
"Acceptable Daily Intake" or ADI.  The RfD is useful as a reference point for gauging
the potential effects of other doses.  Usually, doses that are less than the RfD are not
likely to be associated with any health risks, and are therefore less likely to be  of
regulatory concern.  As the frequency  of exposures exceeding the RfD increases  and as

-------
                                         9
the size of the excess increases, the probability increases that adverse effects may be
observed in a human population.  Nonetheless, a clear conclusion cannot be
categorically drawn that all doses below the RfD are "acceptable" and that  all  doses in
excess of the RfD are "unacceptable."
     In extrapolating non-carcinogen animal test data to humans, EPA uses an
uncertainty factor (formerly known as a safety factor). The uncertainty factor is
based upon professional judgment and ranges from 10 to 1000.
     IRIS maintains files containing reference doses for chronic noncarcinogenic
health effects, and slope factors and unit risks for chronic  exposures to carcinogens.

-------
                                       10
              4.0  EXPOSURE ASSESSMENT AND KEY ASSUMPTIONS

     Once the hazard identification and dose-response assessment have been
completed, the next step is exposure assessment to determine safe exposure levels.
Exposure assessment is the process of characterizing the human populations exposed
to the  chemicals of concern, the environmental transport and fate pathways of those
chemicals, and the frequency, magnitude, and duration of the exposure dose.  In the
water quality criteria methodology, the exposure assessment  includes calculation of the
ambient water concentrations which  do not represent a significant risk to human
health.  The exposure assessments used in calculating the criteria are limited with
respect to accounting for other sources of exposure to a pollutant.  In the case of
carcinogens, exposure assessments assume that all exposure to a pollutant occurs
through ingestion of water and contaminated fish and shellfish and incremental risks
are calculated  on this basis. In the case of non-carcinogens, the equation used has the
ability  to include both intake from other dietary sources and intake from air.
However, if there is insufficient data for air and other dietary sources, it is  assumed
that exposure  is  only from ingestion  of water and fish.
     Ideally, ambient water quality criteria should represent levels for compounds in
ambient water that do not pose a hazard to human populations. However, in any
realistic assessment of human health hazard, it is not possible to attain completely
eliminate hazards or achieve zero risk. Ideally,  criteria would be based on detailed
knowledge of dose-response relationships in humans, including all sources of chemical
exposure, the types of toxic effects elicited, if any, the existence of thresholds for the
toxic effects, the significance of toxicant interactions, and the variances of sensitivities
and exposure levels within human populations.  In practice, such absolute criteria
cannot be established because of deficiencies in both the available data and the means
of interpreting this  information. Consequently, the human health criteria proposed for
carcinogens are designed to minimize, or at least specify the potential risk to humans
due to substances in ambient water.  Human health criteria for non-carcinogens are
designed to minimize the human health hazard.

-------
                                       11
     Potential social or economic costs and benefits are not considered in the
formulation of the criteria. Also, analytical detection limits are not considered in
recommending criteria.
     In the exposure assessment, information is reviewed on current levels of human
exposure to the individual pollutant from all sources.  Much of the data are obtained
from monitoring studies of air, water, food, soil, and human or animal tissue residues.
The major purpose of this review is to provide background information on the
contribution of water exposure relative to all other sources.
     Information on exposure can be valuable in developing and assessing a water
quality criterion.  In these documents exposure from consumption of contaminated
water and contaminated fish and shellfish products is used in criterion formulation.
Data for all modes of exposure are useful in relating total intake to the expected
contribution from contaminated water, fish, and shellfish.  In addition, information on
all routes of exposure,  not limited to drinking water and fish and shellfish ingestion,
can be used to justify or assess the feasibility of the formulation of criteria for
ambient water.
     In development of the criteria, several key assumptions are made and should be
recognized when applying the criteria to  a given situation.  First, the criteria are not
based on water quality monitoring data and, as noted above, the criteria represent
modeled risks, as opposed to  actual risks. The target population is a national average
population.  States may adjust the exposure assumptions to reflect local conditions,
although in practice this has been the exception rather than the rule.
     Second, it is important to recognize that calculation of the criteria does not take
into account special groups of the population which may be more sensitive to the
effects of a particular chemical, such as pregnant  women, young children, or the
elderly. Also, the calculation is explicitly done for a 70 kg adult.  Nor does it take
into account groups which may receive additional exposure to the chemical from other
sources such as occupational exposure. It also does not take into account groups
which may consume large quantities  of fish, such as sport fishermen.

-------
                                       12
     Third, exposure pathways considered in calculating the criteria are ingestion of
drinking water (EPA assumes 2 liters/day contaminated at the criteria concentration),
and consumption of fish (EPA assumes6.5 grams/day, contaminated at the criteria
concentration multiplied by a bioconcentration factor).  The assumed water
consumption of 2 liters/day is taken from the National Academy of Sciences
publication "Drinking Water and Health"  (1977).  The 2 liters/day amount was also
used by EPA in calculating interim drinking water standards  (NAS, 1977). This factor
may be eliminated for water which is not an actual or potential source of drinking
water. In the water quality criteria calculations, an estimated consumption of 6.5
grams/day is assumed for commercially and recreationally harvested fish and shellfish
from estuarine and  fresh waters.  The value  of 6.5 g/day is an average per-capita
consumption rate for  the U.S.  population, including non-consumers (U.S. EPA, 1989).
The inclusion of non-consumers may make the number somewhat under conservative.
As these criteria must serve as national guidance, no attempt is made in the
calculation of the criteria to account for variation among individuals in the amount of
fish consumed. Also, estimates of average U.S. consumption do not account for
subpopulations in areas (such as the Great Lakes) that may consume large quantities
(e.g. 20 g/day) of locally caught sport fish.
     The use of fish consumption  as an exposure factor requires the quantification of
pollutant residues in the edible portions of the ingested species. Accordingly,
bioconcentration factors (BCFs) are used to relate pollutant residues in aquatic
organisms to the pollutant concentration in the ambient waters in which they reside.
Strictly speaking, the  BCF is defined as the theoretical ratio between the concentration
of a chemical in a fish and the concentration of a pollutant in the surrounding water,
at fqirilihriimv Depending on the  mechanisms and rates of chemical transfer (through
the gills, through ingestion, through excretion) a fish may not be at equilibrium and
therefore may exhibit a different ratio. The  term bioaccumulation factor (BAF) refers
to the  measured value of this ratio in a field  situation.  The terms are frequently used
imprecisely and misunderstandings can occur.

-------
                                       13
     One common usage is to consider the BCF only to include intake from water,
and the BAF to additionally include food input. One reason this is misleading is that
BCFs so measured are often measured for short periods of time (to preclude feeding)
and no equilibrium is reached.  These BCFs are, therefore, frequently less than field
measured BAFs.  However, a  true BCF represents  an upper limit which should not be
exceeded.  Should a  fish consume an amount in excess of that determined by the true
BCF, it would simply excrete  the excess or actually serve as a source of the pollutant
to the  water, through the  gills. There have, however, been some studies which
suggest thermodynamic  equilibrium may not always be present in the environment.
     Three different procedures are used to estimate the BCF, depending upon the
lipid solubility of the chemical and the availability of bioconcentration data.  For lipid-
soluble compounds, the  average BCF is calculated from the weighted average  percent
lipids in the edible portions of consumed freshwater and  estuarine fish and shellfish.
This weighted average was calculated to be 3.0 percent using data on consumption of
each species and its corresponding percent lipids.  Because the steady-state BCFs for
lipid-soluble compounds are proportional to percent lipids in tissues,  BCFs for fish and
shellfish can be adjusted to the average percent lipids for aquatic organisms consumed
by the  U.S. population.  For many lipid-soluble pollutants, at least one BCF has  been
determined for which the percent lipid value was  measured in the exposed tissues.
     With 3.0 percent as the weighted average percent lipids for freshwater and
estuarine fish and shellfish in the average diet, a BCF, and a corresponding percent
lipid value, the weighted average bioconcentration factor can be calculated.  For
example, where:

     Weighted average  percent lipids for average  diet =  3.0 %, and
     Measured BCF for Trichloroethylene with bluegills at 4.8% lipids = 17
     Weighted average  BCF for average diet equals:
                            BCF =  17 x 3.0% = 10.6
                                       4.8%

     As an estimate, 10.6 is used for the BCF for the water quality criteria.

-------
                                        14
      In those cases where an appropriate bioconcentration factor is not available, the
equation "Log BCF = (0.85 Log P) - 0.70" has beeen used by EPA to estimate the BCF
for aquatic organisms containing about 7.6 percent lipids from the octanol/water
partition coefficient P.  An adjustment for percent lipids in the average diet versus 7.6
percent is made in order to derive the weighted average bioconcentration factor.
There are other similar equations in general use.
      For non-lipid-soluble compounds, the available  BCFs for the edible portion of
consumed freshwater and estuarine fish and shellfish are weighted according to
consumption factors  for these various types of fish to determine  a weighted BCF
representative of the average diet.
      Human body weight is assumed to be 70 kg in calculations of ambient water
quality criteria.  It should be noted that  this body weight is not  protective of pregnant
women and children. In other risk calculations, EPA has used 50 kg to represent
pregnant women.
      As noted in the previous sections on hazard identification and dose response
relationships, the existing toxicity data must be reviewed to determine the type of
effects of the chemical and the dose response relationship.  If the chemical is classified
as a carcinogen, it is assumed to have no threshold,  and the incremental risks of
developing cancer are calculated.  If the  compound is not classified as a carcinogen, it
is assumed to have a threshold below which effects will not be observed, and the
criterion is calculated based on this RfD.  Examples of calculations for both a
carcinogen and a noncarcinogen are given below.

         4.1  Calculation of Criteria for Non-threshold Effects (Carcinogens^

      After the decision has been made that a compound has the potential for causing
cancers in humans, and that data exist which permit the derivation of a criterion, the
water concentrations which are estimated to cause a lifetime, upper bound
carcinogenic risk of 10s, 10*, and  107 are determined.  The lifetime carcinogenicity

-------
                                       15
risk is the probability that a person would get cancer sometime in his or her life
assuming continuous exposure to the compound.
     The data used for quantitative estimates are of two types: (1)  lifetime animal
studies, and (2)  epidemiologic studies where excess cancer risk has been associated
with exposure to the agent. In animal studies it is assumed, unless  evidence exists to
the contrary, that if a carcinogenic response has been documented at the dose levels
used in the study, then proportionately lower responses will also be observed at all
lower doses, with an incidence determined by a linear extrapolation model which
calculates the carcinogenic  slope factor discussed in the Dose-Response Assessment
Section of this report.
     An example of how criteria are calculated for carcinogens is described below.
Since carcinogens are assumed to have a nonthreshold  dose/response characteristic,
there is no recognized safe concentration for a human carcinogen.  Therefore,  the
recommended concentration of a carcinogen in water for maximum  protection of
human health is zero.  Because attaining a zero concentration level  may not be
feasible in some cases, and in  order to assist EPA and the States in  the possible future
development of water quality regulations, concentrations of the carcinogenic
compound corresponding to several incremental lifetime cancer risk  levels are
estimated.  A cancer risk level provides an estimate of  the additional incidence of
cancer that may be  expected in an exposed population.  A risk of 10* for example,
indicates a  probability of one additional case of cancer for every 100,000 people
exposed; a risk of 10* indicates one additional case of cancer for every million people
exposed, and so forth.
     For carcinogens, the criteria calculation is as follows:

        Criterion =   	fbody weight x risk level)	
                   potency x  (water  consumption + fish consumption x BCF)

     The calculations for hexachlorobutadiene (HCBD) are described here as an
example of criteria formulation for a carcinogen.  HCBD  is produced in the United
States as a by-product of the manufacture of chlorinated hydrocarbons such as

-------
                                        16
tetracholoethylene, trichloroethylene, and carbon tetrachloride.  It is used as a solvent
for many organic substances.  HCBD has a low vapor pressure and, thus, may not
volatilize rapidly from the aqueous environment to the atmosphere. Concentrations
observed in water indicate the HCBD may be quite persistent in the environment.  It
may also be adsorbed onto the sediments, particularly in areas high in organic
content.
      Review of toricity literature indicates the kidney appears to be the organ most
sensitive to HCBD. The  carcinogenic effects of renal tubular adenomas and
adenocarcinomas were strongly demonstrated at a dosage of 20 mg/kg/day in the diet
in rats in a two year feeding study.  Incidence rates for dose  levels of 0.0, 0.2, 2.0,
and 20.0 were 1/90,  0/40, 0/40, and 9/39, respectively.  A q," (carcinogenic slope
factor) value of 0.07752 (mg/kg/day)'1 was calculated from the rat study data.
      Bioconcentration factors are available for  HCBD but the necessary data
concerning percent lipids are not.  Therefore, the equation "Log BCF  = (0.85 Log P) -
0.70" can be used  to  estimate the BCF for aquatic organisms  that contain about 7.6
percent lipids  from the octanol/water partition  coefficient (P).  Based on a measured
log P value of 1.82, the steady-state bioconcentration factor for HCBD is estimated to
be 7.03.  An adjustment  factor of 3.0/7.6 = 0.395 can be used to adjust the
estimated BCF from 7.6 percent lipids (on which the equation is based), to the 3.0
percent lipids, which  is the weighted average for consumed fish and shellfish.  Thus,
the weighted average bioconcentration factor for HCBD and the edible portion of all
freshwater and estuarine aquatic organisms consumed by the  U.S. population is
calculated to be 7.03 x 0.395 = 2.78.
      Parameters used in calculation of the criterion for the 10* risk level are as
follows:

      Body weight  =  70 kg
      Risk level  =  lO*
      Potency  = 0.07752 (mg/kg/day)'1
      Water consumption = 2 liters/day
      Fish consumption = 6.5 grams/day = 0.0065 kg/day
      Bioconcentration factor = 2.78

-------
                                       17
The calculation of an ambient water quality criterion for HCBD based on fish
consumption at the  10"" risk level using the  formula given above is as follows:

Criterion =	(70 kg x IP"6!	
             0.07752 (mg/kg/day)' x (2 liters/day + 0.0065 kg/day x 2.78)

         = 0.45 ug/L

The criteria for the  upper bound risk levels  of 10* and 107 are 4.5 ug/L and 0.045
ug/L, respectively, assuming consumption of 2 liters of drinking water and 6.5 grams
of fish and shellfish per day.  Approximately one percent of the HCBD exposure
results from consumption of aquatic  organisms.  If the exposure is assumed to  be from
fish alone, the water quality criteria  associated with risk levels  of 10*, 10*, and 107
are 500 ug/L,  50 ug/L, and 5.00 ug/L, respectively.

                 4.2  Calculation of Criteria for Threshold Effects

     In developing  guidelines for deriving criteria based on noncarcinogenic responses,
five types of response levels are considered:
     NOEL:     No-Observed-Effect Level
     NOAEL:    No-Observed-Adverse-Effect Level
     LOEL:      Lowest-Observed-Effect Level
     LOAEL:     Lowest-Observed-Adverse-Effect  Level
     PEL:       Frank-Effect Level

Adverse effects are defined  as any effects  which  result in functional impairment and/or
pathological lesions  which may affect the  performance of the whole organism,  or
which reduce an organism's ability to respond to an additional  challenge.  Available
data on the five types of effects are evaluated to determine the threshold effects

-------
                                        18
concentration, and to estimate an RfD.  In determining an RfD, an uncertainty factor
(formerly referred to as a safety factor) ranging from 10 to 1000 is applied to the
estimated threshold level, depending on the confidence in the available data.  When
the quality and quantity of experimental data are satisfactory, a low uncertainty factor
is used; when data are judged to be inadequate or equivocal, a larger uncertainty
factor is used.  In other words, uncertainty factors are assigned on a case specific
basis based on somewhat subjective judgements of the quality and quantity of the
data.  EPA maintains a data base on reference doses (RfDs) in the IRIS system.  The
RfD is an estimate (with uncertainty spanning perhaps an order of magnitude) of a
daily exposure to the human population (including sensitive subgroups) that is likely
to be without appreciable risk of deleterious effects during a lifetime.  RfDs are based
on an assumption of long-term exposure and may not be appropriately applied to
short-term exposure  situations.
      For threshold chemicals  (noncarcinogens), the criteria are based  on an RfD, and
dietary and  inhalation  exposures are considered where data is available. The formula
is:
           Criterion =  	RfD - (dietary intake  + air intake)
                        (water consumption + fish consumption x BCF)
     Toluene is used below as an example for calculation of criteria for a threshold
chemical. Toluene is used in the production of benzene and other chemicals as well
as being used directly in gasoline and as a solvent.  Although it is volatile, it has been
detected in finished water supplies at levels ranging from 0.1 ug/L to 11 ug/L.
     A number of investigations of the subacute and chronic toricity of toluene have
been conducted.  Although most were inhalation studies, at least one long-term oral
dosing study was conducted in which female rats were given toluene at 118, 354, and
590 mg/kg in olive oil by stomach tube five times weekly for 193 days.  No adverse
effects on growth  appearance and behavior, mortality, organ/body weights, blood urea
nitrogen levels, bone  marrow counts, peripheral blood counts, or morphology of major

-------
                                       19
organs were observed at any dose level.  The lack of toxicity reported in this study is
supported by findings of other groups of investigators who found no evidence of
residual injury in a variety of animal species subjected to toluene vapor for varying
times over periods as long as 18 months.  Therefore, the highest dose utilized in the
oral study, 590 mg/kg was used as  the basis for calculating the RfD for toluene.
Since no other data were available on effects at higher levels, 590 mg/kg was
assumed to be the "maximum-no-effect" dose.  However, since the highest dose had  no
observed effect, no information was available on the lowest observed effects.  A single
oral dose of 2.4 g/kg had no hepatotoxic effects in rats and the oral LDM for toluene
in young adult rats was found to be 7.0 g/kg-  It is possible that the actual
"maximum-no-effect" dose may be lower than 590 mg/kg, should alternative indices of
toxicity be evaluated.  Humans may prove to be more sensitive to toluene than
experimental animals.  Thus, a safety factor of 1,000 was applied.  Assuming a body
weight of 70 kg and adjusting for the dosing of five times per week, the RfD is
calculated as follows:
               RfD = 590 mg/kg x 70 kg x 5/7 dav = 29.5 mg/day
                                 1,000
     It is assumed that 100 percent of man's exposure comes from water. Although it
is desireable to arrive at a criterion level for water based upon total exposure
potential, the database for exposures other than water is not sufficient to allow a
factoring of air intake and dietary intake (other than fish and shellfish) into  the
calculation of the criterion.
     No measured steady-state bioconcentration factor (BCF) was available for
toluene.  Therefore, the equation "Log BCF = (0.85 Log P) - 0.70" was used to
estimate  the BCF for aquatic organisms that contain about 7.6% lipids for the octanol
water partition coefficient (P).  Based on an average measured Log P value of 2.51,
the steady-state bioconcentration factor for toluene is estimated to be 27.1.  An
adjustment factor of 3.0/7.6 = 0.395 can be used to adjust the estimated BCF from

-------
                                       20
7.6% lipids (on which the equation is based) to 3.0% lipids, is the weighted average
for consumed fish and shellfish. Thus, the weighted average bioconcentration factor
for toluene and the edible portion of all freshwater and estuarine aquatic organisms
consumed by the U.S. population is calculated to be 27.1 x 0.395 =  10.7
      Consumption of 2 liters of water daily and 6.5 g of contaminated fish having a
bioconcentration factor of 10.7, would result in, assuming 100% gastrointestinal
absorption of toluene, a maximum permissible concentration of 14.3 mg/L for the
ingested water:
                  Criterion =     29.5 mg/dav     = 14.3 mg/L
                               2 L + (10.7 x 0.0065)
Drinking water contributes 97% of the assumed exposure, while eating contaminated
fish products accounts for 3%.  The criterion level for toluene can alternatively be
expressed as 424 mg/L if exposure is assumed to be from the consumption of fish and
shellfish products alone.
     The principal goal of the water quality criteria program is to provide guidance
for maintaining acceptable  water quality under a program for source reduction. The
development of an acceptable level for fish tissue is not a goal of the water quality
criteria methodology, and the SAB did not review the methodology for this use.
However, by taking the criteria and multiplying by the BCF factor used in its
derivation, an implicit fish  tissue level may be calculated.  This  could then be used to
establish a fish contamination advisory, if measured fish levels exceed this level.  This
calculation is the same for  carcinogens and non-carcinogens.  For example, using the
criterion for toluene, the acceptable fish tissue level would be calculated as follows:

    Acceptable fish tissue level = Criterion x BCF = 14.3 ppm x 10.7 =  153 ppm

-------
                                       21
                       5.0  RISK MANAGEMENT DECISION

     Water quality criteria for the protection of human health are issued as guidance
for making risk management decisions. These criteria present scientific data and
guidance on the environmental effects of pollutants which can be used to derive
regulatory requirements for water quality, such as State water quality standards.  The
criteria are EPA's recommended ambient water concentrations which EPA believes do
not represent a significant risk to human health.  States are free to make risk level
determinations as part of their risk management process.
     The water  quality criteria methodology is purely a risk assessment  as opposed to
a risk management process.  No  benefits  analysis is considered, nor are costs.  The
differences  between carcinogens  and non-carcinogens involve the dose-response factors
described above.  For non-carcinogens, the risk assessment leads to a one value
criterion.  For carcinogens, the assessment leads to a series of criteria associated with
incremental increases in the risk of developing cancer. The criteria are calculated for
individual chemicals  with no consideration of additive, synergistic, or antagonistic
effects in mixtures.
     Also,  the criteria do not take into account special populations which may be at
additional risk, such  as sport fisherman or regional populations who consume
quantities of fish much higher than the national average.  If the conditions within a
State differ from the assumptions made in calculation of the criteria, the States have
the options to perform their own risk assessments and to set standards which more
accurately reflect their specific conditions and which provide adequate protection for
human health.

-------
                                      22

                               6.0 REFERENCES


40 Code of Federal Regulations (CFR), Part 136, "Guidelines Establishing Test
     Procedures for the Analysis of Pollutants Under the Clean Water Act".  (49
     Federal Register 2, October 26, 1984).


45 Federal Register 79347, "Water Quality Criteria Documents; Availability,
     Appendix C - Guidelines and Methodology Used in the Preparation of Health
     Effect Assessment Chapters of the Consent Decree Water Criteria Documents"
     November 28, 1980.
National Academy of Sciences (NAS), 1977.  "Drinking Water and Health".
U.S. Environmental Protection Agency (U.S. EPA), 1989.  "Assessing Human
     Health Risks from Chemically Contaminated Fish and Shellfish: A Guidance
     Manual". EPA-503/8-89-002. September.

-------
                  APPENDIX A




WATER QUALITY CRITERIA DOCUMENTS; AVAILABILITY




            (45 Federal Register 79318)

-------
Friday
November 28, 1980
Part V



Environmental

Protection Agency

Water Quality Criteria Documents;
Availability

-------
79318
Federal Register / Vol. 45, No. 231 / Friday, November 28,1080 / Notices
ENVIRONMENTAL PROTECTION
AGENCY
IFRL 1623-3]

Water Quality Criteria Documents;
Availability
AGENCY: Environmental Protection
Agency.
ACTION; Notice of Water Quality Criteria
Documents.   .          	=	

SUMMARY: EPA announces the
availability and provides summaries of
water quality criteria documents for 84
toxic pollutants or pollutant categories.
These criteria are published pursuant to
section 304(a)(l) of the Clean Water Act
AVAILABILITY OP DOCUMENTS:
Summaries of both aquatic-based and
health-based criteria from the
documents are published below. Copies
of the complete documents for
individual pollutants may be obtained
from the National Technical Information
Service  (NTIS), 5285 Port Royal Road.
Springfield. VA 22161, (703-487-4650). A
list of the NTIS publication order
numbers for all 64 criteria documents is
published below. These documents are
also available for public inspection and
copying during normal business hours
at: Public Information Reference Unit,
U.S. Environmental Protection Agency,
Room 2404 (rear], 401M St, S.W.,
Washington, D.C. 20460. As provided La
40 CFR Part 2, a reasonable fee may be
charged for copying services. Copies of
these documents are also available for
review in the EPA Regional Office
libraries.
  Copies of the documents are not
available from the EPA office listed
below. Requests sent to that office will
be forwarded to NTIS or returned to the
sender.                 .           •
  1. Acenaphthene. PB81-117269. -
  2. Acroleln, PB81-117277.
  3. Acrylonltrile, PB81-117285.
  4. Aldrin/Dieldrin, PB81-117301.
  5. Antimony, PB81-117319.
  6. Arsenic, PB81-117327.
  7. Asbestos. PB81-117335.
  a. Benzene, PB81-117293.
  8. Benzldlne, PB81-117343.
  10. Beryllium, PB81-117350.
  11. Cadmium, PB81-117368.
  12. Carbon Tetrachloride, PB81-
117376,
  13. Chlordane, PB81-117384.
  14. Chlorinated benzenes, PB81-
117392.
  15. Chlorinated ethanes. PB81-117400.
  16. Chloroalkyl ethers, PB81-117418.
  17. Chlorinated naphthalene, PB81-
117426.
  18. Chlorinated phenols, PB81-117434.
  18. Chloroform, PB81-117442.
  20.2-chlorophenol, PB81-1174S9,
                      21. Chromium, PB81-117467.
                      22. Copper. PB81-117475.
                      23. Cyanides, PB81-117483.
                      24. DDT, PB81-117491.
                      25. Dichlorobenzenes. PB81-117509.
                      26. Dlchlorobenzidine, PB81-117517.
                      27. Dichloroethylenes. PBB1-117525.
                      28. 2,4-dlchlorophenol, PB81-117533.
                      29. Dichloropropanes/propenes, PB01-
                    117541.
                      30.2,4-dimethylphahol, PBB1-117558,
                     • 31. Dlnltrotoluene, PB81-117S6&
                      32. Diphenylhydrazine, PB81-117731.
                      33. Endosulfan. PB81-117574.
                      34. Endrln, PB81-117582.
                      35. Ethylbenzene, PB81-117890.
                      36. Fluoranthene, PB81-117608.
                      37. Haloethers. PB81-117616.
                      38. Halomethanes, PB81-117624.
                      39. Heptachlor, PB81-117832.
                      4a Hexachlorobutadiene, PB81-
                    117640.
                      41. Hexachlorocyclohexane, PB81-
                    1176B7.
                      42, Hexachlorocyclopentadlene. PB81-
                    117665.
                      43. Isophorone, PB81-117873.
                      44. Lead, PB81-117681.
                      45. Mercury, PB81-117699.
                      48. Naphthalene. PB81-117707.
                      47. Nickel, TB81-117715.
                      48. Nitrobenzene, PB81-117723.
                      49. Nitrophenols, PB81-117749.
                      50. Nitfosamines, PBB1-117758.
                      51. Pentachlorophenol, PB81-117704.
                      52. Phenol PB81-117772.  '
                      53. Phthalate esters. PB81-117780.
                      54. Polychlorinated biphenyls (PCBs),
                    PBai-117788.
                      55. Polynuclear aromatic
                    hydrocarbons, PB81-11780B.
                      56. Selenium, PB81-117814.
                      57. Silver,  PB81-117822.
                      58. Tetrachloroethylene. PB81-117830.
                      59. Thallium, PB81-117848.
                      60. Toluene, PB81-117855.
                      61. Toxaphene. PB81-117863.
                      62. Trichloroethylene, PB81-117871.
                      63. Vinyl chloride, PB81-117889.
                      64. Zinc. PB81-117897.
                    FOR FURTHER INFORMATION CONTACT:
                    Dr. Frank Gostomskl, Criteria and
                    Standards Division (WH-585), United
                    States Environmental Protection
                    Agency, 401M Street, S.W.,
                    Washington, D.C. 20460, [202] 245-3042.
                    SUPPLEMENTARY INFORMATION:

                    Background
                      Pursuant to section 304[a)[l) of the
                    Clean Water Act. 33 U.S.C. 1314(a)(l),
                    EPA is required to periodically review
                    and publish criteria for water quality
                    accurately reflecting the latest scientific
                    knowledge:
                      (A) on the kind and extent of all
                    identifiable affects on health and welfare
                    including, but not limited to, plankton, fish,
shellfish, wlldllfa, plant life, shorelines.
beaches, esthetics, and recreation which may
be expected from the presence of pollutants
in any body of water, Including ground water.
(B) on the concentration and dispersal of
pollutants, or their byproducts, through
biological, physical, and'chemical processes.
and (C) on the effects of pollutants on
biological community diversity, productivity,
and stability, including information on the
factors affecting rates of eutrophicalion and
rates of organic a nd' inorganic sedimentation
for varying types of receiving waters.'  •
  EPA is today announcing the
availability of criteria documents for 64
of the 65 pollutants designated as toxic
under section 307(a)(l) of the Act The
document on TCDD (Dioxin) will be
published within the next month after
review of recent studies. Criteria for the
section 307(a)(l) toxic pollutants being
published today will replace the criteria
for those same pollutants found in the
EPA publication. Quality Criteria far
Water, (the "Red Book.") Criteria for all
other pollutants and,water constituents
found in the "Red Book" remain valid.
The criteria published today have been
derived using revised methodologies for
determining pollutant concentrations
that will, when not exceeded,
reasonably protect human health and
aquatic life. Draft criteria documents
were made available for public
comment (44 FR15926, March 15,1979,
44 FR 43660, July 25,1879, 44 FR 56628,
October 1,1978). These final criteria-
have been derived after consideration of
all comments received.
  These criteria documents are also
issued in satisfaction of die Settlement
Agreement in Natural Resources
Defense Council, et al. v. Train 8 EJtC.
2120 (1976). modified, 12 E.R.C. 1833
(D.D.C. 1979). Pursuant to paragraph 11
of that agreement EPA is required to
publish criteria documents for the 65
pollutants which Congress, in the 1977
amendments to the Act, designated as
toxic under section 307(a)(l). These
documents contain recommended
maximum permissible pollutant
concentrations consistent with the
protection of aquatic organisms, human
health, and some recreational activities.
Although paragraph 11 Imposes certain
obligations on the Agency, it does not
create additional authority.

The Development of Water Quality
Criteria
.  Section 304(a)[l) criteria contain two
essential types of information: (1)
discussions of available scientific data
on the effects of pollutants on public
health and welfare, aquatic life and
recreation, and [2] quantitative
concentrations or qualitative
assessments of the pollutants in water
which will generally ensure water

-------
                   Federal Register  / Vol. 45. No.. 231_ /  Friday.  November 28, 1980 /  Notices
                                                                       79319
 quality adequate to support a specified
 water use. Under section 304(a)(l), these
 criteria are based solely on data and
 scientific judgments on the relationship
 between pollutant concentrations and
 environmental and human health
 effects. Criteria values do not reflect
 considerations of economic or
 technological feasibility.
   Publication of water quality criteria of
 this type has been an ongoing process
 which EPA,  and its predecessor Agency.
 tJle.Federal Water Pollution Control
 Administration, have been engaged in
 since 1968. At that time the first Federal
 compilation of water quality criteria, the
 so-called "Green Book" (Water Quality
 Criteria), was published. As now, these
 criteria contained both narrative
 discussions  of the environmental effects
 of pollutants on a range of possible uses
 and concentrations of pollutants
 necessary to support these uses. Since
 that time, water quality criteria have
 been revised and expanded with
 publication of the "Blue Book" (Water
 Quality Criteria1972) in 1973 and the
 "Red Book"  (Quality Criteria for Water)
 in 1976.
  Since publication of the Red Book
 there have been substantial changes in
 EPA*s approach to assessing scientific
 data and deriving section 304(a)[l]
 criteria. Previous criteria were derived
 from a limit edtlata base. For many
 pollutants, an aquatic life criterion was
 derived by multiplying the lowest
 concentration known to have acute
 lethal effect  on half of a test group of an
 aquatic species (the LC50 value) by an
 application factor in order to protect
 against chronic effects. If data showed a
 substance to be bioaccumulative or to
have other significant long-term effects,
 a factor was used to reduce the
indicated concentrations to a level
presumed to be protective. Criteria for
 the protection of human health were
 similarly derived by considering the
pollutants' acute, chronic, and
bioaccumulative effects on non-human
mammals and humans.
  Although a continuation of the
process of criteria development, the
criteria published today were derived
using revised methodologies
(Guidelines) for calculating the Impact
of pollutants on human health and
 aquatic organisms. These Guidelines
consist of systematic methods for
 assessing valid and appropriate data
 concerning acute and chronic adverse
 effects of pollutants on aquatic
 organisms, non-human mammals, and
 humans. By use of these data in
 prescribed ways, criteria are formulated
 to protect aquatic life and human health
 from exposure to the pollutants. For
 some pollutants, bioconcentration
 properties are used to formulate criteria
 protective of aquatic life uses. For
 almost all of the pollutants,
 bioconcentration properties are used to
 assess the relative extent of human
 exposure to the pollutant either directly
 through ingestion of water or indirectly
 through consumption of aquatic
 organisms. Human health criteria for
 carcinogens are presented as
 incremental risks to man associated   .
 with specific concentrations of the
 pollutant in ambient water. The
 Guidelines used to derive criteria
 protective of aquatic life and human
 health are fully described in appendices
 B and C, respectively, of this Notice.
  The Agency believes that these
 Guidelines provide criteria which more
 accurately reflect the effects of these
 pollutants on human health and on
 aquatic organisms and their uses. They
 are based on a more rational and
 consistent approach for using scientific
 data. These Guidelines were developed
 by EPA scientists in consultation with
 scientists from outside the Agency and
 they have been subjected to intensive
 public comment.
  Neither the Guidelines nor the criteria
 are considered inflexible doctrine. Even
 at this time. EPA is taking action to
 employ the resources of peer review
groups, including the Science Advisory
Board, to evaluate recently published
data, and EPA is conducting its own
evaluation of new data to determine
whether revisions to the  criteria
documents would be warranted.
  The criteria published  today are
based solely on the effect of a single
pollutant. However, pollutants in
combination may have different effects
because of synergistic, additive, or
antagonistic properties. It is impossible
in these documents to quantify the
combined effects of these pollutants.
and persons using criteria should be
aware that site-specific analysis of
actual combinations of pollutants may
be necessary to give more precise
indications of the actual  environmental
impacts of a discharge.
Relationship of the Section 304(a)(l) .
Criteria to Regulatory Programs
  Section 304(a)(l) criteria are not rules
and they have  no regulatory inpact.
Rather, these criteria present scientific
data and guidance on the ehviromental
effect of pollutants which can be useful
to derive regulatory requirements based
on considerations of water quality
impacts. Under the Clean Water Act,
these regulatory requirements may
include the promulgation of water
quality-based effluent limitations under
section 302, water quality standards
under section 303, or toxic pollutant
effluent standards under section 307.
States are encouraged to begin to
modify or, if necessary, develop new
programs necessary to support the
implementation of regulatory controls
for toxic pollutants. As appropriate,
States may incorporate criteria for toxic
pollutants, based on this guidance, into
their water quality standards.
  Section. 304(a)(l) criteria have been
most closely associated with the
development of State water quality
standards/and the "Red Book" values
have; in the past, been the basis for
EPA's assessments of the adequacy of
State requirements. However, EPA is
now completing a major review of its
water quality standards policies and
regulations. After consideration of
comments received on an Advance
Notice of Proposed Rulemaking (43 FR
29588. July 10,1978) and the draft
criteria documents, the Agency intends
to propose, by the end of this year, a
revised water quality standards
regulation which will clarify the
Agency's position on a number of
significant standards issues.
  With the publication of these criteria.
however, it is. appropriate to discuss
EPA's current thinking on standards
issues relating to their use. This
discussion does not establish new
regulatory requirements and is intended
as guidance on the possible uses of
these criteria and an indication of future
rulemaldng the Agency, may undertake.
No substantive requirements will be
established without further opportunity
for public comment.
Water Quality Standards
  Section 303 of the Clean Water Act
provides that water quality standards be
developed for all surface waters. A
water quality standard consists
basically of two parts:' (1) A "designated
use" for which the water body is to be
protected (such as "agricultural,"
"recreation" or "fish and wildlife"), and
(2) "criteria" which are numerical
pollutant concentration limits or
narrative statements necessary to
preserve or achieve the designated use.
A water quality standard is developed
through State or Federal rulemaking
proceedings and must be translated into
enforceable effluent limitations in  a
point source (NPDES) permit or may
form the basis of best management
practices applicable to nonpoint sources
under section 208 of the Act.

Relationship of Section W4(a)(l)
Criteria to the Criteria Component of
State Water Quality Standards:
  In the ANPRM, EPA announced  a
policy of "presumptive applicability" for

-------
 79320
Federal  Register / Vol. 45, No. 231  / Friday, November  28, 1980 / Notices
 section 304(a](l) criteria codified in the
 "Red Book." Presumptive applicability
 meant that a State had to adopt a
 criterion for a particular water quality
 parameter at least as stringent as the
 recommendation in the Red Book unless
 the State was able to justify a less
 stringent criterion based on: natural
 background conditions, more recent
 scientific evidence, or local, site-specific
 information. EPA Is rescinding the
 policy of presumptive applicability
 because it has proven to be too
 inflexible in actual practice.
  Although (he section 304(a)(l) criteria
 represent a reasonable estimate of
 pollutant concentrations consistent with
 the maintenance of designated water
 uses, States may appropriately modify
 these values to reflect local conditions.
 In certain circumstances, the criteria
 may not accurately reflect the toxicity of
 a pollutant because of the effect of local
 water quality characteristics or varying
 sensitivities of local populations. For
 example, in some cases, ecosystem
 adaptation may enable a viable,
 balanced aquatic population to exist in
 waters with high natural background
 levels of certain pollutants. Similarly,
 certain compounds may be more or less
 toxic in some waters because of
 differences in alkalinity, temperature,
 hardness, and other factors.
  Methods for adjusting the section
 304(a)(l) criteria to reflect these local
 differences are  discussed below.
Relationship of Section 304(a)(l)
 Criteria to Designated Water Uses:
  The criteria published today can be
 used to support the designated uses
 which are generally found in State
standards. The  following section
 discusses the relationship between the
criteria and Individual use"
classifications.  Where a water body is
designated for more than one use,  -
criteria necessary to protect the most
sensitive use should be applied.
  1. Recreation: Recreational uses of
water include such activities as
swimming, wading, boating and fishing.
Although insufficient data exist on the
 effects of toxic pollutants resulting from
 exposure through such primary, contact
as swimming, section 304(a)(l) criteria
 based on human health effects may be
 used to support this designated use
where fishing is included in the State
 definition of "recreation." In this
 situation only the portion of the criterion
 based on fish consumption should be
 used.
  2. Protection and Propagation of Fish
 and Other Aquatic Life: The section
 304(a)(l) criteria based on toxicity to
 aquatic life may be  used directly to
 support this designated use.
                       3. Agricultural and Industrial Uses:
                     The section 304(a)(l) criteria were not
                     specifically developed to reflect the
                     impact of pollutants on agricultural and
                     industrial uses. However, the criteria
                     developed for human health and aquatic
                     life are sufficiently stringent to protect
                     these other uses. States may establish
                     criteria specifically designed to protect
                     these uses.
                     .  4. Public, Water Supply: The drinking
                    " water exposure component of the
                     human health effects criteria can apply
                     directly to this use classification or may
                     be appropriately modified depending
                     upon whether the specific water supply
                     system falls within the auspices of the
                     Safe Drinking Water Act's (SDWA)
                     regulatory control, and the type and
                     level of treatment imposed upon the
                     supply before delivery to the consumer.
                     The SDWA controls the presence of
                     toxic pollutants in finished ("end-of-
                     tap") drinking water. A brief description
                     of relevant sections of this Act is
                     necessary to explain how the SDWA
                     will work in conjunction with section
                     304(a)(l) criteria in protecting human
                     health from the effects of toxics due to
                     consumption of water.
                       Pursuant to section 1412 of the SDWA,
                     EPA has promulgated "National Interim
                     Primary Drinking Water Standards" for
                     certain organic and inorganic
                     substances. These standards establish
                     "maximum contaminant levels"
                     ("MCLs") which specify the maximum
                     permissible level of a contaminant in
                     water which may be delivered to a user
                     of a public water system now defined as
                     serving a minimum of 25 people. MCLs
                     are established based on consideration
                     of a range of factors Including not only
                     the health effects of the contaminants
                     but also technological and economic   -•
                     feasibility of the contaminants' removal
                     from the supply. EPA is required to
                     establish revised primary drinking water
                     regulations based on the effects of a
                     contaminant on human health, and
                     include treatment capability, monitoring
                     availability, and costs. Under Section
                     1401(l)(D)(i] of the SDWA. EPA is also
                     allowed to establish the minimum
                     quality criteria for water which may be
                     taken into a public water supply system.
                       Section 304(a)(l) criteria provide
                     estimates of pollutant concentrations
                     protective of human health, but do not
                     consider treatment technology, costs
                     and other feasibility factdrs. The section
                     304(a)(l] criteria also include  fish
                     bioaccumulation and consumption
                     factors  in addition to direct human
                     drinking water intake. These numbers
                     were not developed to serve as "end of
                     tap" drinking water standards, and they
                     have no regulatory significance under
the SDWA. Drinking water standards
are established based on considerations
including technological and economic
feasibility, not relevant to section
304(a)(l) criteria. Section 304(a](i)
criteria may be analogous to the
recommended maximum contaminant
levels (RMCLs) under section
1412(b)(l)(B) of the SDWA in which,
based upon a report from the National
Academy  of Sciences, the Administrator
should set target levels for contaminants
in drinking water at which "no known or
anticipated adverse effects occur and
which allows an adequate margin of
safety". RMCLs do not take treatment.
cost, and other feasibility factors into
consideration. Section 304(a)(l) criteria
are, in  concept, related to the health-
based goals specified in the RMCLs.
Specific mandates of the SDWA such as
the consideration of multi-media
exposure,  as well as different methods
for setting maximum contaminant levels
under the  two Acts, may result in
differences between the two numbers.
  MCLs of the SDWA, where they exist,
control toxic chemicals in finished
drinking water. However, because of
variations In treatment and the fact that
only a relatively small number of MCLs
have been developed, ambient water
criteria may be used by the States as a
supplement to SDWA regulations. States
will have  the option of applying MCLs;
section 304(a)(l) human health effects^
criteria, modified section 304(a)(l)
criteria or controls more stringent than
these three to protect against the effects
of toxic pollutants by Ingestion from
drinking water.
  For untreated drinking water supplies.
States may control toxics In the ambient
water through either use of MCLs (if
they exist for the pollutants of concern),
section 304(a)(l) human health effects
criteria, or a more strigent contaminant
level than the former two options.
  For treated drinking water supplies
serving less than 25 people, States may
choose toxics control through
application of MCLs (if they exist for the
pollutants of concern and are attainable
by the  type of treatment) in the finished
drinking water. States also have the
options to control toxics in the ambient
water by choosing section 304(a)(l.)
criteria, adjusted section 304(a)(l)
criteria resulting from the reduction of
the direct  drinking water exposure
component in the criteria calculation to
the extent that the treatment procedure  .
reduces the level of pollutants, or a more
stringent contaminant level than the
former three options.
  For treated drinking water supplies
serving 25 people or greater. States must
control toxics down to levels at least as
stringent as MCLs (where they exist for

-------
                   Federal Register / Vol. 4S.  No. 231  /  Friday. November 2B. 1980  /  Notices             79321
 the pollutants of concern] In the finished
 drinking water. However, States also
 have the options to control toxics in the
 ambient water by choosing section
 304(a)(l) criteria, adjusted section
 304(a}(l) criteria resulting from the
 reduction of the direct drinking water
 exposure component in the criteria
 calculation to the extent that the
 treatment process reduces the level of
 pollutants, or a more stringent
 contaminant  level-than the former three
 options.
 Inclusion of Specific Pollutants in State
 Standards:           •
  To date, EPA has not required that a
 State address any specific pollutant in
 its standards. Although all States have
 established standards for most
 conventional pollutants, the treatment of
 toxic pollutants has been much less
 extensive. In  the ANPRM, EPA
 suggested a policy under which States
 would be required to address a set of
 pollutants and incorporate specific toxic
 pollutant criteria into  water quality
 standards. If. the State failed to
 incorporate these criteria, EPA would
 promulgate the standards based upon
 these criteria pursuant to section
 303Cc)(4JfB}.
  In the forthcoming proposed revision
 to the water quality standard
 regulations, a significant change in
 policy will be proposed relating to the
 incorporation of certain pollutants in
 State water quality standards. This
 proposal will differ from the proposal
 made in the ANPRM. The ANPRM
 proposed an EPA-published list of
 pollutants for which States would have
 had to develop water quality standards.
 This list might have contained some (or
 all) of the 6! toxic pollutants. However,
 the revised water quality standards
 regulation will propose a process by
 which EPA will assist States in
 identifying specific toxic pollutants
 required for assessment for possible
 inclusion in State water quality
 standards. For these pollutants, States
will have the  option of adopting the
 published criteria or of adjusting those
 criteria based on site-specific analysis.
  These pollutants would generally
represent the greatest  threat to
sustaining a healthy, balanced
ecosystem in water bodies or to human
health due to  exposure directly or
indirectly from water. EPA is currently
developing a process to determine
 which pollutants a State must assess for
possible inclusion in its water quality
standards. Relevant factors might
include the toxicily of the pollutant, the
 frequency and concentration of its
 discharge, its  geographical distribution,
 he breadth of data underlying the
scientific assessment of its aquatic life
and human health effects, and the
technological and economic capacity to
control the discharge of the pollutant.
For some of the pollutants, all States
may be required to assess them for
possible inclusion in their standards. For
others, assessment would be restricted
to States cr limited to specific water
bodies where the pollutants pose a
particular site-specific problem.
Criteria Modification Process •
  Flexibility la available in the
application of these and any other valid
water quality criteria to regulatory
programs. Although in some cases they
may be used by the States as developed,
the criteria may be modified to refect
local environmental conditions and
human exposure patterns before
Incorporation into programs such as
water quality standards. If significant
impacts of site-specific water quality
conditions in the toxicities of pollutants
can be demonstrated or significantly
different exposure patterns of these
pollutants to humans can be shown,
section 2Q4(a)(l) criteria may be
modified to reflect these local
conditions. The term "local" may refer
to any appropriate geographic area
where common aquatic environmental
conditions or exposure patterns exist.
Thus, "local" may signify a Statewide,
regional, river reach, or entire river
basin area. On the other hand, the
criteria of some pollutants might be
applicable nationwide without the need
for adaptation to reflect local
conditions. The degree of toxicity
toward aquatic organisms and humans
characteristic of these pollutants would
not change significantly due to local
water quality conditions,
  EPA is examining a series of
environmental factors or water quality
parameters which might realistically be
expected to affect the laboratory-
derived water quality criterion
recommendation for a specific pollutant.
Factors such as hardness, pH,
Suspended solids, types of aquatic
organisms present, etc. could impact on
the chemical's effect in the aquatic
environment. Therefore, local
information can be assembled and
analyzed to adjust the criterion
recommendation if necessary.
  The Guidelines for deriving criteria for
the protection of aquatic life suggest
several approaches for modifying the
criteria. First, toxicity data, both acute
and chronic, for local species could be
substituted for some or all of the species
used in deriving criteria for the water
quality standard. The minimum data
requirements should still be fulfilled in
calculating a revised criterion. Second,
 criteria may be specifically tailored to a
 local water body by use of data from
 toxicity tests performed with that
 ambient water. A procedure such as this
 would account for local environmental
 conditions in formulating a criterion
 relevant to the local water body. Third,
 site-specific water quality
 characteristics resulting in either
 enhancement or mitigation of aquatic
 life toxicity for the pollutant could be
.• factored into- final formulation of the  •
 criterion. Finally, the criteria may be
 made more stringent to ensure
 protection of an individual species not
 otherwise adequately protected by  any
 of the three modification procedures
 previously mentioned.
   EPA does not intend  to have States
 assess every local stream segment and
 lake in the country on an individual
 basis before determining if an
 adjustment is necessary. Rather, it is
 envisioned that water bodies having
 similar hydrological, chemical, physical,
 and biological properties will be
 grouped for the purpose of criteria
 adjustment. The purpose of this effort is
 to assist States in adapting the section
 304(a) criteria to local conditions where
 needed, thereby precluding the setting of
 arbitrary and perhaps unnecessarily
 stringent or underprotective criteria in a
 water body. In all cases, EPA will still
 be required, pursuant to section 303(c),
 to determine whether the State water
 quality standards are consistent with
 the goals of the Act, including a
 determination of whether State-
 established criteria are adequate to
 support a designated use.
 Criteria for the Protection of Aquatic
 Ofe
 Interpretation of the Criteria
   The aquatic life criteria issued today
 are summarized In Appendix A of this
 Federal Register notice. Criteria have
 been formulated by applying a  set of
 Guidelines to a data base for each
 pollutant The criteria for the protection
 of aquatic life specify pollutant
 concentrations which, if not exceeded,
 should protect most, but not necessarily -
 all, aquatic life and its uses. The
 Guidelines specify that criteria should
 be based on an array of data from
 organisms, both plant and animal,
 occupying various trophic levels. Based
 on these data, criteria can be derived
 which should be adequate to protect the
 types of organisms necessary to support
 an aquatic community.
   The Guidelines are not designed to
 derive criteria which will protect all life
 stages of all species under all
 conditions. Generally some life stage of
 one or more tested species, and

-------
79322
                   Federal  Register / Vol. 45. No. 231  /  Friday. November 28. 1980 / Notices
probably some untested species, will
have sensitivities below the maximum
value or the 24-hour average under some
conditions and would be adversely
affected if the highest allowable
pollutant concentrations and the worst
conditions existed for a long time. In
actual practice, such a situation is not
likely to occur and thus the aquatic
community as a whole will normally be
protected if the criteria are not
exceeded.'In any aquatic community
there is 8 wide range of Individual
species sensitivities to the effects of
toxic pollutants, A criterion adequate to
protect  the most susceptible life stage of
the most sensitive species  would in
many cases be more stringent than
necessary to protect the overall aquatic
community.
  The aquatic life criteria specify both
maximum and 24-hour average values.
The combination of the two values is
designed to provide adequate protection
of aquatic life and its uses from acute
and chronic toxicity and
bioconcentratlon without being as
restrictive as a one-number criterion
would have  to be to provide the same
amount of protection, A time period  of
24 hours was chosen in order to ensure
that concentrations not reach harmful
levels for unacceptably long periods.
Averaging for longer periods, such as a
week or a month for example, could
permit high concentrations to persist  •
long enough to produce significant
adverse effects. A 24-hour  period was
chosen instead of a slightly longer or
shorter period in recognition of daily
fluctuations in waste discharges and of
the  influence of daily cycles of sunlight
and darkness and temperature on both
pollutants and aquatic organisms.
  The maximum value, which is derived
From acute toxicity data, prevents
significant risk of adverse impact to
organisms exposed to concentrations
above the 24-hour average. Merely
specifying the average value over a  '
specified time period is insufficient
because concentrations of chemicals
higher than the average value can kill or
cause irreparable damage in short
periods. Furthermore, for some
chemicals the effect of intermittent high
exposures is cumulative, It is therefore
necessary to place an upper limit on
pollutant concentrations to which
aquatic  organisms might be exposed.
The two-number criterion is intended to
describe the highest average ambient
water concentration which will produce
a water quality generally suited to the
maintenance of aquatic life while
restricting the extent and duration of the
excursions over that average to levels
which will not cause harm. The only
 way to assure the same degree of
 protection with a one-number criterion
 would be to use the 24-hour average as a
 concentration that is not to be exceeded
 at any time in any place.
   Since some substances may be more
 toxic in freshwater than in saltwater, or
 vice versa, provision is made for
 deriving separate water quality criteria
 for freshwater and for saltwater for each
 substance. However, for .some
 substances sufficient data may not be
 available to derive one or both of these
 criteria using the Guidelines.
   Specific aquatic life criteria have not
 been developed for all of the 65 toxic
 pollutants. In those cases where there
 were insufficient data to allow the
 derivation of a criterion, narrative
 descriptions  of apparent threshold levels
 for acute and/or chronic effects based
 on the available data are presented.
 These descriptions are intended to
 convey a sense of the degree of toxicity
 of the pollutant in the absence of a
 criterion recommendation.
 Summary of the Aquatic Life Guidelines
   The Guidelines for Deriving Water
 Quality Criteria for the Protection of
 Aquatic Life ana its Uses were
 developed to describe an objective,
 internally consistent, and appropriate
 way of ensuring that water quality
 criteria for aquatic life would provide,
 on the average, a reasonable amount of
 protection without an unreasonable
 amount of overprotection or
 underprotection. The resulting criteria
 are not intended to provide 100 percent
 protection of all species and all uses of
 aquatic life all of the time, but they are
 intended to protect most species in a
 balanced, healthy aquatic community.
 The Guidelines are published as
 Appendix B of this Notice. Responses to
 public comments on these Guidelines
 are attached as Appendix 0.
   Minimum data requirements are
 identified in four areas: acute toxicity to
 animals (eight data points), chronic
 toxicity to animals (three data points).
 toxicity to plants, and residues.
 Guidance is also given for discarding
 poor quality  data.
   Data on acute toxicity are needed for
 a variety of fish and invertebrate
 species and are used to derive a Final
 Acute Value. By taking into account the
 number and relative sensitivities of the
 tested species, the Final Acute Value is
 designed to protect most, but not
 necessarily all, of the tested and
 untested species.
"  Data on chronic toxicity to animals
 can be used to derive a Final Chronic
 Value by two different means. If chronic
 values are available for a specified
 number and  array of species, a final
chronic value can be calculated directly.
If not, an acute-chronic ratio is derived
and then used with the Final Acute
Value to obtain the Final Chronic Value
  The Final Plant Value is obtained by
selecting the lowest plant toxicity value
based on measured concentrations.
  The Final Residue Value is intended
to protect wildlife which consume
aquatic organisms and the marketability
of aquatic organisms. Protection of. the •
marketability of aquatic organisms is, in
actuality, protection of a use of that
water body ("commercial fishery"). Two
kinds of data are necessary to calculate
the Final Residue Value: a
bioconcentration factor (BCF) and a
maximum permissible tissue
concentration, which can be an FDA
action level or can be the result of a
chronic wildlife feeding study. For lipid
soluble pollutants, the BCF is
normalized for percent Upids and then
the Final Residue Value is calculated by
dividing the maximum permissible
tissue concentration by the normalized
BCF and by an appropriate percent lipid
value, BCFs are normalized for percent
lipids since the BCF measured for any
individual aquatic species is generally
proportional to the percent lipids in that
species.
  If sufficient data are available to
demonstrate that one or more of the
final values should be related to a water
quality characteristic, such as salinity,
hardness, or suspended solids, the final
value(s) are expressed as a function of
that characteristic.
  After the four final values (Final
Acute Value, Final Chronic Value. Final
Plant Value, and Final Residue Value)
have been obtained, the criterion is
established with the Final Acute Value
becoming the maximum value and the
lowest oft he other three values  .
becoming the 24-hour average value. All
of the data used to calculate the four
final values and any additional pertinent
information are then reviewed to
determine if the criterion is reasonable.
If sound scientific evidence indicates
that  the criterion should be raised or
lowered, appropriate changes are made
as necessary.
  The present Guidelines have been
revised from the earlier published
versions (43 FR 21B08, May 18,1978; 43
FR 28028, July 8,1878; 44 FR 1S926,
March 15,1979). Details have been
added in many places and the concept
of a minimum data base has been
incorporated. In addition, three
adjustment factors and the species
sensitivity factor have been deleted.
These modifications were the result of
the Agency's analysis of public
comments and comments received from
the Science Advisory Board on earlier

-------
                   Federal  Register / Vol. 45. No. 231  /  Friday.  November 28. 1980 / Notices             79323
 versions of the Guidelines. These
 comments and the Resultant
 modifications are addressed fully in
 Appendix D to this notice.
 Criteria for the Protection of Human
 Health

 Intespretation of the Human Health
 Criteria       ,
   The human health criteria issued
 today are summarized in Appendix A of
. this Federal Register notice. Criteria for
 the protection of human Health are
 presented for 62 of the 85 pollutants
 based on their carcinogenic, toxic, or
 organoleptic (taste and odor) properties.
 The meanings and practical uses of the
 criteria values are distinctly different
 depending on the properties on which
 they are based.
   The objective of the health
 assessment portions of the criteria
 documents is to estimate ambient water
 concentrations which, in the case of
 non-carcinogens, prevent adverse health
 effects in humans, and in the case of
 suspect or proven carcinogens, represent
 various levels of incremental cancer
 risk.
   Health assessments typically contain
 discussions of four elements: Exposure,
 phannacokinetics, toxic effects, and
 criterion formulation.
   The exposure section summarizes
 information on exposure routes:
 ingestion directly from water, indirectly
 from consumption of aquatic organisms
 found in ambient water, other dietary
 sources, inhalation, and dermal, contact.
 Exposure assumptions are used to
 derive human health criteria. Most
 criteria are based solely on  exposure
 from consumption of water containing a
 specified concentration of a toxic
 pollutant and through consumption of
 aquatic organisms which are assumed to
 have bioconcentrated pollutants from
 the water in which they live. Other
 multimedia routes of exposure such as
 air, non-aquatic diet, or dermal are not
 factored into the criterion formulation
 for the vast majority of pollutants due to
 lack of data. The criteria are calculated
 using the combined aquatic  exposure
 pathway and also using the  aquatic
 organism ingestion exposure route
 alone. In criteria reflecting both the
 water consumption and aquatic
 organism ingestlon routes of exposure,
 the relative exposure contribution varies
 with the propensity of a pollutant to
 hue-concentrate, with the consumption of
 aquatic organisms becoming more
 important as the bioconcentration factor
 (BCF) increases. As additional
 information on total exposure is
 assembled for pollutants for which
 criteria reflect only the two specified
 aquatic exposure routes, adjustments in
 water concentration values may be
 made. The Agency intends to publish
 guidance which will permit the States to
 identify significantly different exposure
 patterns for their populations. If
 warranted by the demonstration of
 significantly different exposure patterns,
 this will become an element of a process
 to adapt/modify human health-based
 criteria to local conditions, somewhat  '
 analogous to the aquatic life criteria •
 modification process discussed
 previously. It is anticipated that States
 at their discretion will be able to set
 appropriate human health criteria based
 on this process.
  The phannacokinetics section reviews
 data on absorption, distribution,
 metabolism, and excretion to assess the
 biochemical fate of the compounds in
 the human and animal system. The toxic
 effects section reviews data on acute,
 subacute. and chronic toxicity,
 synergistic and antagonistic effects, and
 specific information on mutagenicity,
 teratogenicity. and carcinogenicity.
 From this review, the toxic effect to be
 protected against is identified taking
 into account the quality, quantity, and
 weight of evidence characteristic of the
 data. The criterion formulation section
 reviews the highlights of the text and
 specifies a rationale for criterion
 development and the mathematical
 derivation of the criterion number.
  Within the limitations of time and
 resources, current published information
 of significance was Incorporated into the
 human health assessments. Review
 articles and reports were used for data
 evaluation and synthesis. Scientific
 judgment was exercised in reviewing
 and evaluating the data in each criteria
 document and in identifying the adverse
 effects for which protective criteria were
 published,
~ Specific health-based criteria are
 developed only if a weight of evidence
 supports the occurrence of the toxic
 effect and if dose/response data exist
 from which criteria can be-estimated.
  Criteria for suspect or proven
 carcinogens are presented as
 concentrations in water associated with
 a range of incremental cancer risks to
 man. Criteria for non-carcinogens
 represent levels at which exposure to a
 single chemical is not anticipated to
 produce adverse effects in man. In a few
 cases, organoleptic (taste and odor) data
 form the basis for the criterion. While
 this type of criterion does not represent
 a value which directly affects human
 health, it is presented as an estimate of
 the level of a pollutant that will not
 produce unpleasant taste or odor either
 directly from water consumption or
 indirectly by consumption of aquatic
organisms found in ambient waters. A
criterion developed in this manner is
judged to be as useful as other types of
criteria in protecting designated water
uses. In addition, where data are
available, toxicity-based criteria are
also presented for pollutants with
derived organoleptic criteria. The choice
of criteria used in water quality
standards for these pollutants will
depend upon the designated use to be
protected. In the" case of a multiple use
water body, the criterion protecting the
most sensitive use will be applied.
Finally, for several pollutants no criteria
are recommended due to a lack of
information sufficient for quantitative
criterion formulation.
Risk Extrapolation •
  Because methods do not now exist to
establish the presence of a threshold for
carcinogenic effects, EPA's policy is that
there is no scientific basis for estimating
"safe" levels for carcinogens. The
criteria for carcinogens, therefore, state
that the recommended concentration for
maximum protection of human health is
zero. In addition, the Agency has
presented a range of concentrations
corresponding to incremental cancer
risks of 10"f to 10" * (one additional case
of cancer in populations ranging from
ten million to 100,000, respectively).
Other concentrations representing
different risk levels may be calculated
by use of the Guidelines. The risk
estimate range is presented for
Information purposes and does not
represent an Agency Judgment on an
"acceptable" risk level.
Summary of the Human Health
Guidelines
  The health assessments and
corresponding criteria published today
were derived based on Guidelines and
Methodology Used in the Preparation of
Health Effect Assessment Chapters of
the Consent Decree Water Criteria
Documents (the Guidelines) developed
by EPA's Office of Reserch and
Development. The estimation of health
risks associated with human exposure to
environmental pollutants requires
predicting the effect of low doses for up
to a lifetime in duration. A combination
of epldemiological and animal dose/
response data is considered the
preferred basis for quantitative criterion
derivation. The complete Guidelines are
presented as Appendix C. Major issues
associated with these Guidelines and
responses to public comments are
presented as Appendix E.
  No-effect (non-carcinogen) or   .
specified risk (carcinogen)
concentrations were estimated by
extrapolation from animal toxicity or

-------
 79324
Federal  Register / Vol. 45. No. 231  /  Friday. November 28. 1980 / Notices
human epidemiology studies using the
following basic exposure assumptions: a
70-kilogram male person (Report of the
Task Croup on Reference Man,
International Commission for Radiation
Protection. November 23,1957) as the
exposed individual; the average daily
consumption of freshwater and
estuarine fish and shellfish products
equal to 6.5 grams/day; and the  average.
ingestiott of two liters /day of water
(Drinking Water and Health, National
Academy of Sciences, National
Research Council, 1977). Criteria based
on these assumptions are estimated to
be protective'of an adult male who
experiences average exposure
conditions.
  Two basic methods were used to
formulate health criteria, depending on
whether the prominent adverse effect
was cancer or other toxic
manifestations. The following sections
detail these methods,
Carcinogens
  Extrapolation of cancer responses
from high to low doses and subsequent
risk estimation from animal data is
performed using a linearized multi-stage
model. This procedure is flexible enough
to fit all monotonlcally-increasing dose
response data, since it incorporates
several adjustable parameters. The
multi-stage model Is a linear non-
threshold model as was the* "one-hit"
model originally used in the proposed
criteria documents. The linearized multi-
stage model and its characteristics are
described fully in Appendix C. The
linear non-threshold concept has been
endorsed by the four agencies in the
Interagency Regulatory Liaison Group
and is less likely to underestimate risk
at the low dosestypical of
environmental exposure than .other
models that could be used. Because of
the uncertainties associated with dose
response, animal-to-human
extrapolation and other unknown
factors, because of the use of average
exposure assumptions, and because of
the serious public health consequences
that could result if risk were
underestimated, EPA believes that it is
prudent to use conservative methods to
estimate risk  in the water quality
criteria program. The linearized
multistage model Is more systematic  and
invokes fewer arbitrary assumptions
than the "one-hit" procedure previously
used.
  It should be noted that extrapolation
models provide estimates of risk since a
varitey of assumptions are built.into  any
model. Models using widely different
assumptions may produce estimates
ranging over several orders of
magnitude. Since Ihere is at present no
                     way to demonstrate the scientific
                     validity of any model, the use of risk
                     extrapolation models Is a subject of
                     debate in the scientific community.
                     However, risk extrapolation Is generally
                     recognized as .the only tool available at
                     this time for estimating the magnitude of
                     health hazards associated with non-
                     threshold toxicants and has been
                     endorsed by numerous Federal agencies ,
                     and scientific organizations, including
                     EPA's Carcinogen Assessment Croup,
                     the National Academy of Sciences, and
                     the Interagency Regulatory Liaison
                     Group as a useful means of assessing
                     the risks of exposure to various
                     carcinogenic pollutants.
                     Non-Carcinogens
                       Health criteria based on toxic effects
                     of pollutants other than carcinogenlcity
                     are estimates of concentrations which
                     are not expected to produce adverse
                     effects in humans. They are based upon
                     Acceptable Dally Intake (ADI) levels
                     and are generally derived using no-
                     observed-adverse-effect-level (NOAELJ
                     data from animal studies although
                     human data are used wherever
                     available. The ADI is calculated using
                     safety factors to account for
                     uncertainties inherent in extrapolation
                     from animal to man. In accordance with
                     the National Research Council
                     recommendations//W/7A7/i# Water and
                     Health. National Academy of Sciences,
                     National Research Council, 1977), safety
                     factors of 10,100, or 1,000 are  used
                     depending on the quality and  quantity of
                     data. In some instances extrapolations
                     are made from inhalation studies or
                     limits to approximate a human response
                     from Ingestion using the Stokinger-
                     Woodward model (Journal of American
                     Water Works Association, 1958).
                     Calculations of criteria from ADIs are
                     made using the standard exposure
                     assumptions {2 liters of water. 6.5 grams
                     of edible aquatic products, and an
                     average body weight of 70 kg].
                       Dated: October 24,1980.
                     Douglas M. Costle,
                     Administrator,

                     Appendix A—Summary of Water
                     Quality Criteria
                     Acenaphthene
                     Fresh water A qua tic Life
                       The available data for acenaphthene
                     indicate that acute toxicity to freshwater
                     aquatic life occurs at concentrations as
                     low as 1,700 fig/1  and would occur at
                     lower concentrations among species
                     that are more sensitive than those
                     tested. No  data are available concerning
                     the chronic toxicity of acenaphthene to
                     sensitive freshwater aquatic animals but
 toxicity to freshwater algae occur at
 concentrations as low as 520 fig/1.

 Saltwater Aquatic Life        »

  The available data for acenaphthene
 indicate that acute and chronic toxicity
 to saltwater aquatic life occur at
 concentrations as low as 970 and 710
 fig/1, respectively, and would occur at
 lower concentrations among species
 that are more sensitive than those  •
 tested. Toxicity to algae occurs at
 concentrations as low as 500 jig/1.

 Human Health
  Sufficient data is not available for
 acenaphthene to derive a level which
 would protect against the potential
 toxicity of this compound. Using
 available organoleptic data, for
 controlling undesirable taste and odor
 quality of ambient water, the estimated
 level is 20 fig/1. It should be recognized
 that organoleptic data as a basis for
 establishing a water quality criteria
 have limitations and have no
 demonstrated relationship to potential
 adverse human health effects.'.

 Acrolein                   ;;

 Freshwater Aquatic Life    ',•'
  The* available data for acrolein
 indicate that acute and chronic toxicitjr
 to freshwater aquatic life occurs at  •• •
 concentrations as low as 68 and 21 fig/1,
 respectively, and would occur at lower
 concentrations among species that are
 more sensitive than those tested.

 Saltwater Aquatic Life
  The available data for acrolein
 indicate that acute toxicity to saltwater
 aquatic life occurs at concentrations as
 low as 55 fig/1 and would occur at lower
 concentrations among species that are
 more sensitive than those tested. No
 data are available concerning the
 chronic toxicity of acrolein to sensitive
 saltwater aquatic life.

 Human Health         .  •'
  For the protection of human health
 from the toxic properties of acrolein
 ingested through water and
 contaminated aquatic organisms, the
 ambient water criterion is determined to
 be 320 ug/1.
  For the protection of human health
 from the toxic properties of acrolein
 ingested through contaminated aquatic
 organisms alone, the ambient water
.criterion is determined to be 780 fig/I.

 Acrylonitrile

 Freshwater Aquatic Life
  The available data for acrylonitrile
 indicate that acute toxicity to freshwater
 aquatic life occurs at concentrations as

-------
                  Federal Register / Vol.  45, No. 231 / Friday,  November 28,  1980 / Notices
                                                                      79325
low as 7,550 fig/I and would occur at
lower concentrations among species
that are more sensitive than those
tested. No definitive data are available
concerning the chronic toxicity of
acrylonitrile to sensitive freshwater
aquatic life but mortality occurs at
concentrations as low as 2,600 jig/1 with
a fish species exposed for 30 days.

Saltwater Aquatic Life
  .Only, one saltwater species has been
tested with acrylonitrile and no
statement can be made concerning acute
or chronic toxicity.  \

Human Health
  For the maximum protection of human
health from the potential carcinogenic
effects due lo exposure of acrylonitrile
through ingestion of contaminated water
and contaminated aquatic organisms,
the ambient water concentration should
be zero based on the non-threshold
assumption for this chemical. However,
zero,level may not be attainable at the  .
present time. Therefore, the levels which
may result In incremental increase of
cancer risk over the lifetime are
estimated at 10~8, 10"«, and 1CT'. The
corresponding criteria are .58 ftg/1, -058
Kg/1 and .006 pg/1, respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the
levels are 8J j*g/l, .65 jig/1 and ,065 pgl
L respectively. Other concentrations
repreaenting'different risk levels may be
calculated by use of the Guidelines. The
risk estimate range is presented for
information purposes and does not
represent an Agency judgment on an
"acceptable" risk level.

Aldrin-Dieldrin

DieldrJn
Freshwater Aquatic Life
  For dieldrin the criterion to protect
fresh water aquatic life as derived using
the Guidelines is 0.0019 p.g/1 as a 24-
hour average and the concentration
should not exceed 2.5 jig/1 at any time.
Saltwater Aquatic Life
  For dieldrin the criterion to protect
saltwater aquatic life as derived using
the Guidelines is 0.0019 jtg/1 as a 24-
hour average and the concentration
should not exceed 0.71 j»g/l at any time.
Human Health
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of dieldrin
through ingestion of contaminated water
and contaminated aquatic organisms,
the ambient water concentration should
be zero based on the non-threshold
assumption for this chemical. However,
zero level may not be attainable at the
present time. Therefore, the levels which
may result in incremental increase of
cancer risk over the lifetime are
estimated at 10"», 10"*, and 10''. The
corresponding criteria are .71 ng/l, .071
ng/l. and .0071 ng/l, respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the  .
levels are .78 ng/l, .076 ng/l, and .0076
ng/l respectively. Other concentrations
representing different risk levels may be
calculated by use of the Guidelines, The
risk estimate range is presented for
Information purposes and does not
represent an Agency judgment on an
"acceptable" risk level.

Aldrin
Fresh water Aquatic Life

  For freshwater aquatic life the
concentration of aldrin should not
exceed 9.0 jig/1 at any time. No data are
available concerning the chronic toxicity
of aldrin to sensitive freshwater aquatic
life.

Saltwater Aquatic Life

  For saltwater aquatic life the
concentration of aldrin should not
exceed 1.3 pg/1 at any time. No data are
available concerning the chronic toxicity
of aldrin to sensitive saltwater aquatic
life.

Human Health

  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of aldrin through
ingestion of contaminated water and
contaminated aquatic organisms, the
ambient water concentration should be
zero based on the non-threshold
assumption for this chemical. However,
zero level may not be attainable at the
present time. Therefore, the levels which
may result in incremental increase of
cancer risk over the lifetime are
estimated at 10"*, 10"*, and 10~*. The
corresponding criteria are .74 ng/l, .074
ng/l. and .0074 ng/l, respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water,  the
levels are .79 ng/l, .079 ng/l, and .0079
ng/l, respectively. Other concentrations
respresenting different risk levels may
be calculated by use of the Guidelines.
The risk estimate range is presented for
information purposes and does not
represent an Agency judgment on an
"acceptable" risk level.
Antimony
Freshwater Aquatic Life
  The available data for antimony
indicate that acute and chronic toxicity
to freshwater aquatic life occur at
concentrations as low as 9.000 and 1.600
ug/1. respectively, and would occur at
lower concentrations among species
that are more sensitive than those
tested. Toxicity to algae occurs at
concentrations as* low as<610 {*§/!•

Saltwater Aquatic Life
  No saltwater organisms have been
adequately tested with antimony, and
no statement can be made concerning
acute or chronic toxicity.

Human Health
  For the protection of human health
from the toxic properties of antimony
ingested through water and
contaminated aquatic organisms, the
ambient water criterion is determined to
be 148 fig/1,
  For the protection of human health
from the toxic properties of antimony
ingested through contaminated aquatic
organisms alone, the ambient water
criterion is determined to be 45,000 fig/I.

Arsenic
Freshwater Aquatic Life
  For freshwater aquatic life the
concentration of total recoverable
trlvalent inorganic arsenic should not
exceed 440 pg/1 at any time. Short-term
effects on embryos and larvae of aquatic
vertebrate species have been shown to
occur at concentrations as low as 40 jig/
1.

Saltwater Aquatic Life
  The available data for total
recoverable trivalent inorganic arsenic
indicate that acute toxicity to saltwater
aquatic life occurs at concentrations as
low as 508 fig/' and would occur at
lower concentrations among species
that are more sensitive than those
tested. No data are available concerning
the chronic toxicity of trivalent
Inorganic arsenic to sensitive saltwater
aquatic life.
Human Health
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of arsenic
through ingestion of contaminated water
and contaminated aquatic organisms,
the ambient water concentration should
be zero based on the non-threshold
assumption for this chemical. However,
zero level may not be attainable  at the
present time. Therefore, the levels which
may result in incremental increase of
cancer risk over the lifetime are

-------
 79326
Federal  Register / Vol. 45. No. 231  /  Friday. November 2B. 1980  /  Notices
 estimated at 1   •,:
 chronic toxicity,    ,       .         '•

 Human Health             ?      ; ,; ^
   For the maximum protection of human
 health from the potential carqinogenic;
 effects due to exposure  of beryllium
 through ingestion of contaminated wafer
 and contaminated aquatic organisms,
 the ambient water concentration should
 be zero based on the non-threshold
 assumption for this chemical. However,
 zero level may not be attainable at the
 present time. Therefore, the levels which
 may result in incremental increase of
 cancer risk over the lifetime are
 estimated at 10"8,10'*, and 10"'. The
 corresponding criteria are 37 ng/1, 3.7
 ng/1, and .37 ng/1. respectively. If the
 above estimates are made for
 consumption of aquatic organisms only,
 excluding consumption  of water, the
 levels are 641 ng/1,84.1 ng/1, and 6,41
 ng/1, respectively. Other concentrations
 representing different risk levels may be
 calculated by use of the Guidelines. The
' risk estimate range is presented for
 information purposes and does not
 represent an Agency judgment on an
 "acceptable" risk level.
 Cadmium

 Freshwater Aquatic Life
   For total recoverable cadmium the
 criterion (in M8/1) to protect freshwater
 aquatic life as derived using the
 Guidelines is the numerical value given

-------
                  Federal Register / Vol. 45. No. 231 / Friday. November 28. 1980 / Notices
                                                                     79327
                  w „„ „ 24_nour
 average and the concentration (in fig/1)
 should not exceed the numerical value
 given by e" » U«««*H«H-»-I* at any
 time. For example, a hardnesses of 50,
 ICO, and 200 mg/1 as CaCO, the criteria
 are 0.012,0.025, and 0.051 fig/1.
 respectively, and the concentration of
 total recoverable cadmium should not
 exceed 1.5, 3.0 and 6.3 pg/i, respectively,
 at any time.

 Saltwater Aquatic Life
  For total recoverable cadmium the
 criterion to protect saltwater aquatic life
 as derived using the Guidelines is 4.5
 fig/1 as a 24-hour average and the
 concentration should not exceed 59 fig/1
 at any time.

 Human Health
 . The ambient water quality criterion
 for cadmium is recommended to be
 identical to the existing drinking water
 standard which is 10 fig/L Analysis of
 the toxic effects data resulted In a
 calculated level which is protective of
 human health against the ingestion of
 contaminated .water and contaminated
 aquatic organisms. Hie calculated value
 is comparable to the present standard.
 For this reason a selective criterion
 based on exposure solely from
 consumption of 6.5 grams of aquatic
 organisms was not derived-
 Carbon Tetrachloride

 Freshwater Aquatic Life
  The available date foir carbon
 tetrachloride indicate that acute toxicity
 to freshwater aquatic life occurs at
 concentrations as low as 35,200 fig/1 and
would occur at lower concentrations
among species  that are more sensitive
than those tested. No data are available
concerning the chronic toxicity of
carbon tetrachloride to sensitive
 freshwater aquatic life.

Saltwater Aquatic Life  ;
  The available data for carbon
 tetrachloride indicate that acute toxicity
 to saltwater aquatic life occurs at
concentrations as low as 60,000 ftg/1 and
would occur at lower concentrations
 among species  that are more sensitive
that those tested. No data are available
 concerning the  chronic toxicity of
carbon tetrachloride to sensitive
saltwater aquatic life.

Human Health
  For the maximum protection of human
health from the potential  carcinogenic
effects due to exposure of carbon
tetrachloride through ingestion  of
contaminated water and contaminated
aquatic organisms the ambient water
concentration should be zero based on
the non-threshold assumption for this  -
chemical. However, zero level may not
be attainable at the present time.
Therefore, the levels which may result in
incremental Increase of cancer risk over
the lifetime are estimated at 10"*, 10~*..
and 10~T. The corresponding criteria are
4.0ftg/l, .40f»g/I, and.04 fig/1,
respectively. If the above estimates are
made for consumption of aquatic
organisms only, excluding consumption
of water, the levels are 8&4 fig/L8.94 • •
fig/L and .69 fig/L respectively. Other
concentrations representing different
risk levels may be calculated by use of
the Guidelines. The risk estimate range
is presented for information purposes
and does not represent an Agency
Judgment on an "acceptable" risk level.
Chlordane              ,
Freshwater Aquatic Ufe
  For chlordane the criterion to protect
freshwater aquatic  life as derived using
the Guidelines la 0.0043 fig/I as a 24-
hour average and the concentration
should not exceed 2.4 fig/1 at any time.
Saltwater Aquatic Ufe
  For chlordane the criterion to protect
saltwater aquatic life as derived using
the Guidelines is 0.0040 fig/1 as a 24-
hour average and the concentration
should not exceed 0.09 fig/1 at any time.
Human Health
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of chlordane
through ingestlon of contaminated water
and contaminated aquatic organisms,
the ambient water concentration should
be zero based on the non-threshold
assumption for this chemical. However,
zero level may not be attainable at the
present time. Therefore,; the levels which
may result in Incremental increase of
cancer risk over the lifetime are
estimated at lO'VlO^, and .10'f. The
corresponding criteria are 4.6 ng/L ,46
ng/1, and .046 ng/L respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the
levels are 4J8 rig/1, .48 ng/1, and .046 ng/
1, respectively. Other concentrations
representing different risk levels may be
calculated by use of the Guidelines. The
risk estimate range  is presented for
information purposes and does not
represent an Agency Judgment on an
"acceptable" risk level.
Chlorinated Benzenes
Freshwater Aquatic Life
  The available data for chlorinated
benzenes indicate that acute toxicity to
freshwater aquatic life occurs at
concentrations as low as 250 fig/1 and
would occur at lower concentrations
among species that are more sensitive
than those tested. No data are available
concerning the chronic toxicity of the
more toxic of the chlorinated benzenes
to sensitive freshwater aquatic life but
toxicity occurs at concentrations as low
as 50 fig/1 for a fish species exposed for
7.5 days.
Saltwater Aquatic L/fe
  The available data for chlorinated
benzenes Indicate that acute and
chronic toxicity to saltwater aquatic life
occur at concentrations as low as 160
and 129 fig/1, respectively, and would
occur at lower concentrations  among
species that are more sensitive than
those tested.

Human Health    •
.  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of '  • .
hexachlorobenzene through ingestion of
contaminated water and contaminated
aquatic:organisms, the ambient water
concentration should be zero based on
the non-threshold assumption  for this
chemical. However, zero level may not
be attainable at the present  time.
Therefore, the levels which may result In
Incremental increase of cancer risk over
the lifetime are estimated at 10"', 10'*,
and icr*. The corresponding
recommended criteria are 7.2 ng/1, .72
ng/1, and .072 ng/1. respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the
levels are 7.4 ng/1, .74 ng/1, and .074 ng/
1, respectively.
  For the protection of human health
from the toxic properties of 1,2,4,5-
tetrachlorobenzene ingested through
water and contaminated'aquatic
organisms, the ambient water criterion
is determined to be 38 fig/L
  For the protection of human  health
from the toxic properties of 1,2,4,5-
tetrachlorobenzene ingested through
contaminated aquatic organisms alone,
the ambient water criterion is
determined to be 48 fig/1..
  For the protection of human  health
from the toxic properties of
pentachlorobenzene ingested through
water and contaminated aquatic
organisms, the ambient water  criterion
is determined to be 74 fig/1.
  For the protection of human health
from the toxic properties of
pentachlorobenzene ingested through
contaminated aquatic organisms alone,
the ambient water criterion is
determined to be 85 ftg/1.
  Using the present guidelines, a
satisfactory criterion cannot be derived

-------
 78328
Federal  Register / Vol. 45. No.  231 / Friday. November 28.  19flQ / Notices
 at this time due to the insufficiency in
 the available data for trichlorobenzene.
  For comparison purposes, two
 approaches were used to derive
 criterion levels for monochlorobenzene.
 Based on available toxicity data, for the
 protection of public health, the derived
 level is 488 MS/1- Using available
 organoleptic data, for controlling
 undesirable taste and odor quality of
 ambient  water, the estimated level is 20
 fig/1. It should be recognized that
 organoleptic data as a basis for
 establishing a water quality criteria
 have limitations and have no
 demonstrated relationship to potential
 adverse human health effects.

 Chlorinated Ethanes

 Freshwater Aquatic Life
  The available freshwater data for
 chlorinated ethanes indicate that
 toxicity increases greatly with
 increasing chlorinetion, and that acute
 toxicity occurs at concentrations as low
 as 118,000 fig/1 for 1,2-dichloroethane.
 18.000 fig/1 for two trichloroethanes.
 9,320 fig/1 for two tetrachloroethanes,
 7,240 fig/1 for pentachloroe thane, and
 980 fig/1 for hexachloroe thane. Chronic
 toxicity occurs at concentrations as low
 as 20.000 fig/1 for 1,2-dichloroethane,
 9,400 M8/I for 1.1.2-trichloroethane, 2,400
 fig/1 for 1.1,2.2,-tetrachloroethane, 1.100
 jig/1 for pentachloroethane. and 540 fig/I
 for hexachloroethane. Acute and
 chronic toxicity would occur at lower
 concentrations among species 'that are
 more sensitive than those tested.
Saltwater Aquatic Life
  The available saltwater data for
chlorinated ethanes Indicate that
 toxicity Increases greatly with
increasing chlorination and that acute
 toxicity to fish and invertebrate species
 occurs at concentrations as low as  ~
113,000 fig/1 for 1.2-dichloroethane,
 31,200 fig/1 for 1,1.1-trichloroe thane,
6.020 fig/1 for 1,1,2,2-tetrachloroethane,
 300 fig/1 for pentachloroethane, and 940
 fig/1 for hexachloroethane. Chronic
 toxicity occurs at concentrations as low
 as 281 fig/1 for pentachloroethane. Acute
 and chronic toxicity would occur at
 lower concentrations among species
 that are more sensitive than those
tested.

Human Health
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of 1,2-di-
 chloroethane through ingestion of
 contaminated water and contaminated
 aquatic organisms, the ambient water
 concentration should be zero based on
 tile non-threshold assumption for this
                     chemical. However, zero level may not
                     be attainable at the present time.
                     Therefore, the levels which may result in
                     incremental increase of cancer risk over
                     the lifetime are estimated at 10~s, 10"*,
                     and 10"*. The corresponding criteria are
                     9.4 fig/1. .94 fig/1, and .094 fig/1,
                     respectively. If the above estimates are
                     made for consumption of aquatic
                    . organisms only, excluding consumption
                     of water, the levels are 2,430 fig/1, 243
                     ftg/1. and 24.3 fig/1 respectively. Other
                     concentrations representing different
                     risk levels may be calculated by use of
                     the Guidelines. The risk estimate range
                     Is presented for information purposes
                     and does not represent an Agency
                     judgment on an "acceptable" risk level.
                       For the protection of human health
                     from the toxic properties of 1,1,1-
                     trichloroethane  ingested through water
                     and contaminated aquatic organism, the
                     ambient water criterion is determined to
                     bel8.4mg/l.
                       For the protection of human health
                     from the toxic properties of l.l.l.-tri-
                     chloroethane ingested through
                     contaminated aquatic organisms alone,
                     the ambient water criterion is
                     determined to be 1*03 g/L
                       For the maximum protection of human
                     health from the potential carcinogenic
                     effects due to exposure of 1,1,2,-
                     trichloroethane  through ingeslion of *
                     contaminated water and contaminated
                     aquatic organisms, the ambient water
                     concentration should be zero based on
                     the non-threshold assumption; for this
                     chemical. However, zero level may not
                     be- attainable at the present time.:
                     Therefore, the levels which may result in
                     incremental increase of cancer risk over
                     the lifetime are estimated atlO~*. 10'6.
                     and 10"'. The corresponding criteria are
                     6.0 fig/1. ,B fig/1, and .06 fig/1,
                     respectively. If the above estimates are
                     made for consumption of aquatic
                     organisms only, excluding consumption
                     of water, the levels are 418 fig/1,41.8
                     fig/1, and 4J8 fig/1 respectively. Other
                     concentrations representing different
                     risk levels may be calculated by use of
                     the Guidelines. The risk estimate range
                     is presented for information purposes
                     and does not represent an Agency
                     judgment on an "acceptable" risk level.
                       For the maximum protection of human
                     health from the  potential carcinogenic
                     effects due to exposure of 1,1,2,2-tetra-
                     chloroethane through Ingestion of
                     contaminated water and contaminated
                     aquatic organisms, the ambient water
                     concentration should be zero based on
                     the non-threshold assumption for this
                     chemical. However, zero level may not
                     be attainable at the present time.
                     Therefore, the levels which may result in
                     incremental increase of cancer risk over
                     the lifetime are  estimated at 10**, 10"*,
and 10~*. The corresponding criteria are
1.7 fig/1, .17 fig/1, and .017 ftg/l,
respectively. Ifthe above estimates are
made for consumption of aquatic
organisms only, excluding consumption
of water, the levels are 107 fig/1,10.7
ftg/1. and 1.07 fig/1, respectively. Other
concentrations representing different
risk levels may be calculated by use of
the Guidelines, The risk estimate range
is presented for information purposes,
and does not represent an Agency
Judgment on an "acceptable" risk level.
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of hexa-
chloroethane through ingestion of
contaminated water and contaminated
aquatic organisms, the ambient water
concentration should be zero based on
the non-threshold assumption for this
chemical. However, zero level may not
be attainable at the present time.
Therefore, the levels which, may result in
incremental increase of cancer risk over
the lifetime are estimated at 10**, 10"*,
and 10" *. The corresponding criteria are
19 fig/L 1.8 MS/1. and ,li fig/1.
respectively. If the above estimates are   ,
made for consumption of aquatic       /'
organisms only, excluding consumption ,
of water, the levels are 87.4 fig/1, 8.74
fig/1; and .87 fig/I, respectively. Other
concentrations representing different
risk levels may be calculated by use of
the Guidelines, The risk estimate range
is presented for information purposes
and does not represent an Agency      -:
judgment on an "acceptable" risk level. ()
  Using the present guidelines, a
satisfactory criterion cannot be derived "
at this time due to the insufficiency in
the available data for
monochloroethane.
  Using the present guidelines, a
satisfactory criterion cannot be derived
at this time due to the insufficiency in
the available data for 1,1,-
dichloroethane.
  Using the present guidelines, a
satisfactory criterion cannot be derived
at this time due to the insufficiency in
the available data for 1.1,1,2-
tetrachloroe thane.
  Using the present guidelines, a
satisfactory criterion cannot be derived
at this time due to the insufficiency in
the available data for
pentachloroethane.

Chlorinated Naphthalenes

Freshwater Aquatic Life
  The available data for chlorinated
naphthalenes Indicate that acute
toxicity to freshwater aquatic life occurs
at concentrations as low as 1,600 ftg/1
and would occur at lower
concentrations among species that are

-------
                   Federal Register / Vol.  45.  No. 231  / Friday, November 28. 1980 / Notices^
                                                                        79329
 more sensitive than those tested. No
 data are available concerning the
 chronic toxicity of chlorinated
 naphthalenes to sensitive freshwater
 aquatic life,

 Saltwater Aquatic Life
   The available data for chlorinated
 napthalenes indicate that acute toxicity
 to saltwater aquatic life occurs at
 concentrations as low as 7.5 ^g/1 and
 would occur at lower, concentrations
 among species that are more sensitive
 than those tested No data are available
 concerning the chronic toxicity of
 chlorinated naphthalenes to sensitive
 saltwater aquatic life.
 Human Health
   Using the present guidelines, a
 satisfactory criterion cannot be derived
 at this time due to the insufficiency in
 the available data for chlorinated
 napthalenes.
 Chlorinated Phenols

 Freshwater Aquatic Life
  .The available freshwater data for
 chlorinated phenols indicate that
 toxicity generally increases with
.increasing chlorineUon, and that acute
 toxicity occurs at concentrations as low
 as 30 jig/1 for 4-chloro-3-methylphenol to
 greater than 500,000 pg/1 for other
 compounds. Chronic toxicity occurs at
 concentrations as low as 970 pg/l for
 2,4,6-trichlorophenol. Acute and chronic
 toxicity would occur at lower
 concentrations among species that are
 more sensitive than those tested.
 Saltwater Aquatic Life
   The available saltwater data for
 chlorinated phenols indicate that
 toxicity generally increases with
 increasing chlorination and that acute
 toxicity occurs at concentrations as low
 as 440 ^g/1 for 2,3,5,6-tetrachlorophenol
 and 29,700 jtg/1 for 4-chlorophenol.
 Acute toxicity would occur at lower
 concentrations among species that are
 more sensitive than those tested. No
 data are available concerning the
 chronic toxicity of chlorinated phenols
 to sensitive saltwater aquatic life.

 Human Health
   Sufficient data is not available for 3-
 nonochJorophenol to derive a level
 •vhich would protect against the
 )otential toxicity of this compound.
 Jsing available organoleptic data, for
 on trolling undesirable taste and odor
 [uality of ambient water, the estimated
 eve! is 0.1  fig/1. It should be recognized
 hat organoleptic data as a basis for
 stablishing a water quality criteria
 ave limitations and have no
demonstrated relationship to potential
adverse human health effects.
  Sufficient data is not available for 4-
monochlorophenol to derive a level
which would protect against the
potential toxicity of this compound.
Using available organoleptic data, for
controlling undesirable taste and odor
quality of ambient water, the estimated
level is 0.1 ftg/1. It should be recognized
that organoleptic data as a basis for
establishing a water quality criteria.
have limitations and have no
demonstrated relationship to potential
adverse human health effects.
  Sufficient data is not available for 2,3-
dichlorophenol to derive a level which
would protect against the potential
toxicity of this compound. Using
available organoleptic data, for •
controlling undesirable taste and odor
quality of ambient water, the estimated
level is .04 j*g/l. It should be recognized
that organoleptic data as a basis for
establishing a water quality criteria
have limitations and have no
demonstrated relationship to potential
adverse human health effects.
  Sufficient data is not available for 2,5-
dichlorophenol to derive a level which
would protect against the potential
toxicity of this compound. Using
available organoleptic data, for
controlling undesirable taste and odor
quality of ambient water, the estimated
level is .5 pg/1. It should be recognized
that organoleptic data as a basis for
establishing a water quality criteria
have limitations  and have no
demonstrated relationship to potential
adverse human health effects.
  Sufficient data is not available for 2,6-
dichlorophenol to derive a level which
would protect against the potential
toxicity of this compound. Using
available organoleptic data, for
controlling undesirable taste and odor
quality of ambient water, the estimated
level is .2 jig/1. It should be recognized
that organoleptic data as a basis for
establishing a water quality criteria
have limitations and have no
demonstrated relationship to potential
adverse human health effects.
  Sufficient data is not available for 3.4-
dichlorophenol to derive a level which
would protect against the potential
toxicity of this compound. Using
available organoleptic data, for
controlling undesirable taste and odor
quality of ambient water, the estimated
level is .3 ftg/1. It should be recognized
that organoleptic data as a basis for
establishing a water quality criteria
have limitations  and have no
demonstrated relationship to potential
adverse human health effects.
  Sufficient data is not available for
2,3,4.6-tetrachlorophenoI to derive a
level which would protect against the
potential toxicity of this compound.
Using available organoleptic data, for
controlling undesirable taste and odor
quality of ambient water, the estimated
level is 1 j*g/l. It should be recognized
that organoleptic data as a basis for
establishing a water quality criteria
have limitations and have no
demonstrated relationship to potential
adverse human health effects.
  For comparison purposes, two •
approaches were used to derive
criterion levels for 2,4,5-trichloropheriol.
Based on available toxicity data, for the
protection of public health, the derived
level is 2.6 mg/1. Using available
organoleptic data, for controlling
undesirable taste and odor quality of
ambient water, the estimated level is 1.0
fig/1. It should be recognized that
organoleptic data as a basis for
establishing a water quality criteria
have limitations and have no
demonstrated relationship to potential
adverse human health effects.
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of 2,4,6-
trichlorophenol through ingestioh of
contaminated water and contaminated
aquatic organisms, the ambient water
concentration should be zero based on
the non-threshold assumption for this
chemical. However, zero level may not
be attainable at the present time.
Therefore, the levels which may result in
incremental increase of cancer risk over
the lifetime are estimated at 10~s, 10~9,
and 10"'. The corresponding criteria are
12 fig/1,1.2 ng/1, and .12 fig/1
respectively. If the above estimates are
made for consumption of aquatic
organisms only, excluding consumption
of water, the levels are 36 Mg/1,3,6 fig/1,
and .36 fig/l« respectively. Other
concentrations representing different
risk levels may be calculated by use of
the Guidelines. The risk estimate range
is presented for information purposes
and does not represent an Agency
judgment on an "acceptable" risk level.
  Using available organoleptic data, for
controlling undesirable taste and odor
quality of ambient water, the estimated
level is 2 fig/1. It should be recognized
that organoleptic data as a basis for
establishing a water quality criterion
have limitations and have no
demonstrated relationship to potential
adverse human health effects.
  Sufficient data is not available for 2-
rnethyl-4-chlorophenol to derive a level
which would protect against any
potential toxicity of this compound.
Using available organoleptic data, for
controlling undesirable taste and odor
quality of ambient water, the estimated
level is 1800 fig/I, It should be

-------
 79330
Federal Register / Vol. 45, No.  231 / Friday, November 28. 1980 / Notices
 recognized that organoleptic data as a
 basis for establishing a water quality
 criterion have limitations and have no
 demonstrated relationship to potential
 adverse human health effects.
   Sufficient data Is not available for 3-
 methyl-4-chlorophenol to derive a level
 which would protect against the
 potential toxiclty of this compound.
 Using available organoleptic data, for
•controlling undesirable taste and odor
 quality of ambient water, the estimated
 level is 3000 jig/I. It should be
 recognized that organoleptic data as a
 basis for establishing a water quality
 criterion have limitations and have no
 demonstrated relationship to potential
 adverse human health effects.
   Sufficient data is not available for 3-
 rnethyl-6-chlorophenol to derive a level
 which would protect against the
 potential toxicity of this compound.
 Using available organoleptic data, for
 controlling undesirable taste and odor
 quality of ambient water, the estimated
 level is 20 fig/1. It should be recognized
 that organoleptic data as a basis for
 establishing a water quality criterion
 have limitations and have no
 demonstrated relationship to potential
 adverse human health effects.
 Chloroalkyl Ethers

 Freshwater Aquatic Ufa
.   The available data for chloraalkyl
 ethers indicate that acute toxicity to
 freshwater aquatic life occurs at
 concentrations as low as 238.000 fig/1
 and would occur at lower
 concentrations among species that BIB
 more sensitive than those tested. No
 definitive data are available concerning
 the chronic toxicity of chloroalkyl ethers
 to sensitive freshwater aquatic life.'

 Saltwater A guatic Life
   No saltwater organisms have been
 tested with any chloroalkyl  ether and no
 statement can be made concerning acute
 and chronic toxicity.

 Human Health
   For the maximum protection of human
 health from the potential carcinogenic
 effects due to exposure of bis-
 (chloromethyl)-ether through Ingestion
 of contaminated water and
 contaminated aquatic organisms, the
 ambient water concentration should  be
 zero based on the non-threshold
 assumption for this chemical. However,
 zero level may not be attainable at the
 present time. Therefore, the levels which
 may result in incremental increase of
 cancer risk over the lifetime are
 estimated at 10"*, 10-«, and 10"'. The
 corresponding criteria are ,038 ng/1,
 .0038 ng/1. and .00038 ng/1. respectively.
                     If the above estimates are made for
                     consumption of aquatic organisms only,
                     excluding consumption of water, the
                     levels are 18.4 ng/1,1.84 ng/1, and .184
                     ng/1, respectively. Other concentrations
                     representing different risk levels may be
                     calculated by use of the Guidelines. The
                     risk estimate range is presented for
                     information purposes and does not
                     represent an Agency judgment on an
                     '"acceptable" risk level.
                       For the maximum protection of human
                     health from the potential carcinogenic
                     effects due to exposure of bis (2-
                     chloroethyl) ether through Ingestion of
                     contaminated water and contaminated
                     aquatic organisms, the ambient water
                     concentration should be zero based on
                     the non-threshold assumption for this
                     chemical. However, zero level may not
                     be attainable at the present time.
                     Therefore, the levels which may result in
                     incremental Increase of cancer risk over
                     the lifetime are estimated at 10'5,10"«,
                     and 10~\ The  corresponding criteria are
                     .3 fig/1, .03 fig/1, and .003 fig/1,
                     respectively. If the above estimates are
                     made for consumption of aquatic
                     organisms only, excluding consumption
                     of water, the levels are 13.8 fig/1,1.38
                     >ig/], and .138 fig/1, respectively; Other
                     concentrations representing different
                     risk levels may be calculated by use of
                     the Guidelines. The risk estimate range
                     is presented for information purposes
                     and does not represent an Agency
                     judgment on an "acceptable" risk level
                       For the protection of human health
                     from the toxic properties of bis  {2-
                     chloroisopropyl) ether ingested through
                     water and contaminated aquatic
                     organisms,  the ambient water criterion
                     is determined to be 34.7 fig/1.
                       For the protection of human health
                     from the toxic properties of bis  {2-
                     chloroisopropyl) ether ingested through
                     contaminated aquatic organisms alone,
                     the ambient water criterion is
                     determined to be 4136 mg/1.

                     Chloroform

                     Freshwater Aquatic Life

                       The available data for choloroform
                     indicate that acute toxicity to freshwater
                     aquatic life occurs at concentrations as
                     low as 28.MO fig/1, and would occur at
                     lower concentrations among species
                     that are more sensitive than the three
                     tested species. Twenty-seven-day LC50
                     values indicate that chronic toxicity
                     occurs at concentrations as low as 1,240
                     fig/1, and could occur at lower
                     concentrations among species or other
                     life stages that are more sensitive than
                     the earliest life cycle stage of the
                     rainbow trout.
Saltwater Aquatic Life

  The data base for saltwater sp|^es is
limited to one test and no statement can
be made concerning acute or chronic
toxicity.
Human Heallh

  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of chloroform
through ingestlon of contaminated water
and contaminated aquatic organisms,
the ambient water concentration should
be zero based on the non-threshold
assumption for this chemical. However,
zero level may not be attainable at the
present time. Therefore, the levels which
may result  in incremental increase of
cancer risk over the lifetime are
estimated at 10"*. 10~*. and lO"*. The
corresponding criteria are 1.90 fig/1, .19
fig/1, and .019 fig/1, respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the
levels are 157 jig/1,15.7 fig/1, and 1.57
fig/1. respectively. Other concentrations
representing different risk levels may be
calculated by use of the Guidelines. The
risk estimate range is presented for
information purposes and does not
represent an Agency judgment on an
"acceptable" risk level.
2-Chlorophenol

Freshwater Aquatic Life

  The availabe data for 2-chlorophenol
Indicate that acute toxicity to freshwater
aquatic life occurs at concentrations as
low as 4,380 fig/1 and would occur at
lower concentrations among species
that are more sensitive that those tested.
No definitive data are available
concerning the chronic toxicity of 2- .
chlorophenol to sensitive freshwater
aquatic life but flavor impairment occurs
in one species of fish at concentrations
as low as 2,000 ng/1.
Saltwater Aquatic Life

  No saltwater organisms have been
tested with 2-chlorophenol and no
statement can be made concerning acute
and chronic toxicity.
Human Health

  Sufficient data is not available for 2-
chlorophenol to derive a level which
would protect against the potential
toxicity of this compound. Using
available organoleptic data, for
controlling undesirable taste and odor
quality of ambient water, the estimated
level is 0.1  fig/1. It should be recognized
that organoleptic data as a basis for
establishing a water quality criteria
have limitations and have no

-------
                  Federal Register / Vol. 45. No.  231 / Friday, November 28, 1980 /Notices
                                                                       79331
demonstrated relationship to potential
adverse human health effects.
Chromium

Freshwater Aquatic Life
  For total recoverable hexavalent
chromium the criterion to protect
freshwater aquatic life as derived using
the Guidelines is 0.29 ftg/1 as a 24-hour
average and the concentration should
not exceed 21 pg/1 at any time.
  For freshwater aquatic life the
concentration (in /ig/1) of total
recoverable trivalent chromium should
not exceed the numerical value given by
'•e(1.08[ln(h8rdness}]+3.48)" at any
time. For example, at hardnesses of 60,
100 and 200 mg/1 as CaCO, the
concentration of total recoverable
trivalent chromium should not exceed
2,200,4,700. and 9.900 MS/1, respectively.
at any time. The available data indicate
hat chronic toxicity to freshwater
iquatic life occurs at concentrations as
ow a 44 jig/1 and would occur at lower
Concentrations among species that are
nore sensitive than those tested.
tali water. Aquatic Life
  For total recoverable hexavalent
Chromium the criterion to protect
altwater aquatic life as derived using
he Guidelines is 18 Mg/1 as a 24-hour
verage and the concentration should
ot exceed 1,260 Mg/1  at any time.
  For total recoverable trivalent
hromium, the availabe data indicate
lat acute toxicity to saltwater aquatic
 fe occurs at concentrations as low as
3,300 M8/1, and would occur at lower
ancentrations amoung species that are
 iore sensitive than those tested. No
 3ta are available concerning the
 ironic toxicity of trivalent chromium to
 '.nsitive saltwater aquatic life.

 umanHealth                    •
  For the protection of human health
 om the, toxic properties of Chromium
 [ ingested through water and   -
 mtamlnated aquatic organisms, the
 nbient water criterion is determined to
.'slTOtr.s/1;:   ;
  For    protection of human health
  imt   .oxic properties of Chromium
  ingested through contaminated
  uatic organisms alone, the ambient
  iter criterion is determined to be 3433
 •i/i.   - ^;vv,-   ;....•:.;:-';-.:
  The ambient water quality criterion
  • total Chromium VI is recommended
  be identical to the existing drinking
  iter standard which is SO jig/1.
  lalysis of the toxic effects data
  •ulled in a calculated level which is
  Hective of human health against the
  estion of contaminated-water and
  itnminated aquatic organisms. The
 calculated value is comparable to the
 present standard. For this reason a
 selective criterion based on exposure
 solely from consumption of 6.5 grams of
 aquatic organisms was not derived.
 Copper

 Fresh water Aquatic Life
   For total recoverable copper the
 criterion to protect freshwater aquatic
 life as derived using the Guidelines is 5.6
 ^g/1 as a 24-hour average and the '   •
 concentration (in pg/1) should not
 exceed the numerical value given by
 e(0.94[ln(hardnes8}]-1.23) at any time.
 For example, at hardnesses of 50,100.
 and 200 mg/1 CaCO, the concentration
 of total recoverable copper should not
 exceed 12,22, and 43 /*g/l at any time.
 Saltwater Aquatic Life
   For total recoverable copper the
 criterion to protect saltwater aquatic life
 as derived using the Guideline! is 4.0
 jig/1 as a 24-hour average and the
 concentration should not exceed 23 j*g/l
 at any time.
 Human Health
   Sufficient data is not available for
 copper to derive a level which would
 protect against the potential toxicity of
 this .compound. Using available
 organoleptic data,  for controlling
 undesirable taste and odor quality of
 ambient  water, the estimated level is 1
 mg/1. It should be recognized that
 organoleptic data as a basis for
 establishing a water quality criteria
 have limitations and have no
 demonstrated relationship to potential
 adverse human health effects.
'Cyanide-  '    ,'•   ..'. .-.
 FreshwaterAquaticLife
  For free cyanide (sum of cyanide
 present as HCN and CN~, expressed as
 CM] the criterion to protect freshwater
 aquatic life as derived using the
 Guidelines is 3.6 jtg/1 as a 24-hour
 average and the concentration should
 not exceed 52 jtg/1 at any time.
 Saltwater Aquatic Life
  The available "data for free cyanide
 (sum  of cyanide present as HCN and
 CN~t  expressed as CN) Indicate that
 acute toxicity to saltwater aquatic life
 occurs at concentrations as .low as 30
 Mg/1 and would occur at lower
 concentrations among species that are
 more sensitive than those tested. If the
 acute-chronic ratio for saltwater
 organisms is similar to that for ,
 freshwater organisms, chronic toxicity
 would occur at concentrations as low as
 2.0 Mg/1 for the tested species and at
 lower concentrations among species
 that are more sensitive than those
 tested.
 Human Health
  The ambient water quality criterion
 for cyanide is recommended to be
 identical to the existing drinking water
 standard which is 200 jig/1. Analysis of
 the toxic effects data resulted in a
 calculated level which is protective of
 human health against the ingestion of
 contaminated water and contaminated
•aquatic organisms. The calculated value
 is comparable to the present standard.
 For this reason a selective criterion
 based on exposure solely from
 consumption of 6.5 grams of aquatic
 organisms was not derived.
 DDT and Metabolites
 Freshwater Aquatic Life     :

 DDT
  For DDT and its metabolites the
 criterion to protect freshwater aquatic
 life as derived using the Guidelines is
 0.0010 M8/1 as a 24-hour average and the
 concentration should not exceed 1.1 M8/1
 at any time.
 TDE
  The available data for TDE indicate
 that acute toxicity to freshwater aquatic
 life occurs at concentrations as low as
 0.6 Mg/1 and would occur at lower
 concentrations among species that are
 more sensitive than those tested. No
 data are available concerning the
 chronic toxicity of TDE to sensitive
 freshwater aquatic life.
 DDE
  t The available data for DDE indicate
 that acute'toxicity to freshwater aquatic
 life occurs at concentrations as low aa
 1.050 jxg/1 and would occur at lower
 concentrations among species that are
 more sensitive than those tested. No
 data are available concerning the
 chronic toxicity of DDE to sensitive
 freshwater aquatic life.
 Saltwater Aquatic Life            '
 DDT
  For DDT and its metabolites the
 criterion to protect  saltwater aquatic life
 as derived using the Guidelines  is 0.0010
 Mg/1 as a 24-hour average and the
 concentration should not exceed 0.13
 Mg/1 at any time.

 TDE
  The available data for TDE indicate
 that acute toxicity to saltwater aquatic
 life occurs at concentrations as low as
 3.6 Mg/1 and would occur at lower
 concentrations among species that are
 more sensitive than those tested. No
 data are available concerning the

-------
79332
Federal Register / Vol. 45. No. 231  / Friday. November 28. 19BO / Notices
chronic toxicity of TDE to sensitive
saltwater aquatic life.

DDE
  The available data for DDE indicate
that acute toxicity to saltwater aquatic
life occurs at concentrations as low as
14 n?/l and would occur at lower
concentrations among species that are
mere  sensitive than those tested. No
data are available concerning the
chronic toxicity of DDE to sensitive
saltwater aquatic life.

Human Health
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of DDT through
Ingestion of contaminated water and
contaminated aquatic organisms, the
ambient water concentration should be
zero based on the non-threshold
assumption for this chemical. However,
zero level may not be attainable at the
present time. Therefore, the levels which
may result in incremental increase of
cancer risk over the lifetime are
estimated at 10~», W*. and 10"*. The
corresponding criteria are .24 ng/1, .024
ng/1. and .0024 ng/1, respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the
levels are .24 ng/l, .024 ng/1, and .0024
ng/1, respectively. Other concentrations
representing different risk levels may be
calculated by use of the Guidelines. The
risk estimate range is presented for
information purposes and does not
represent an Agency Judgment of an
"acceptable" risk level.
Dichiorobenzenes

Freshwater Aquatic Life
  The available data for
dichiorobenzenes indicate that acute
and chronic toxicity to freshwater
aquatic life occurs at concentrations as
low as 1,120 and 763 fig/1, respectively,
and would occur at lower
concentrations among species that are
more  sensitive than those tested.

Saltwater Aquatic Life
  The available data for
dichiorobenzenes indicate that acute
toxicity to saltwater aquatic life occurs
at concentrations as low as 1,970 jig/1
and would occur at lower
concentrations among species that are
more  sensitive than those tested. No
data are available concerning the
chronic toxicity of dichiorobenzenes to
sensitive saltwater aquatic life.
Human Health
  For the protection of human health
from the toxic properties  of
dichiorobenzenes (all isomers) ingested
                     through water and contaminated aquatic
                     organisms, the ambient water criterion
                     is determined to be 400 Mg/1-
                       For the protection of human health
                     from the toxic properties of
                     dichiorobenzenes (all isomers) ingested
                     through contaminated aquatic organisms
                     alone, the ambient water criterion is
                     determined to be 2.8 mg/1.

                    •Dichlorobenzidines..

                     Freshwater Aquatic Life
                       The data base available for
                     dichlorobenzidines and freshwater
                     organisms is limited to one test on
                     bioconcer.tration of 3,3'-
                     dichlorobenzidine and no statement can
                     be made concerning acute or chronic
                     toxicity.
                     Saltwater Aquatic Life

                       No saltwater organisms have been
                     tested with any dichlorobenzidine and
                     no statement can be made concerning
                     acute or chronic toxicity.

                     Human Health

                       For the maximum protection of human
                     health from the potential carcinogenic
                     effects due to exposure of
                     dichlorobenzidine through ingestion of
                     contaminated water and contaminated
                     aquatic organisms,  the ambient water
                     concentration should be zero base on
                     the non-threshold assumption for this
                     chemical. However, zero level may not
                     be attainable at the present time.
                     Therefore, the levels which may result in
                     incremental Increase of cancer risk over
                     the lifetime are estimated at 10""*, 10~«,
                     and 10"'. The corresponding criteria are
                     .103 ftg/1, .0103 fig/1, and .00103 jig/1,
                     respectively. If the above estimates are
                     made for consumption of aquatic
                     organisms only, excluding consumption
                     of water, the levels are .204 ng/1, .0204
                     fig/1, and .00204 fig/1,  respectively.
                     Other concentrations  representing
                     different risk levels may be calculated
                     by use of the Guidelines. The risk
                     estimate range is presented for   .
                     information purposes  and does not
                     represent an Agency Judgment on an
                     "acceptable" risk level.

                     Dichloroethylenes

                     Freshwater Aquatic Life

                       The available data  for
                     dichloroethylenes indicate that acute
                     toxicity to freshwater aquatic life occurs
                     at concentrations as low as 11,600 fig/1
                     and would occur at lower
                     concentrations among speciea that are
                     more sensitive than those tested. No
                     definitive data are available concerning
                     the chronic toxicity of dichlorethylenes
                     to sensitive freshwater aquatic life.
Saltwater Aquatic Life
  The available data for
dichlorethylenes indicate that acute
toxicity to saltwater aqua tic life occurs
at concentrations as low as 224,000 /ig/1
and would occur at lower
concentrations among species that are
more sensitive than those tested. No
data are available concerning the
chronic toxicity dichioroethyleneB to
sensitive saltwater aquatic life.

Human Health
  For the maximum protection of human
health from the potential carcinogenic
effects duo to exposure of
1,1-dichloroethylene through ingestion of
contaminated water and contaminated
aquatic organisms, the ambient water
concentration should be zero bastd on
the non-threshold assumption for this
chemical. However, zero level may not
be attainable at the present time.
Therefore, the levels which may result In
incremental increase of cancer risk over
the lifetime are estimated at 10"'. 10'*,
and 10"'. The corresponding criteria-are
.03 Mg/1. -033 Mg/1, and .0033 Mg/1,
respectively. If the above estimates are
made for consumption of aquatic
organisms only, excluding consumption
of water, the levels are 18.5 pg/1, 1.0S
MS/1, and .185 fig/1, respectively. Other
concentrations representing different
risk levels may be calculated by use of
the Guidelines. The risk estimate range
is presented for information purposes
and does not represent an Agency
Judgment on an "acceptable" risk level,
  Using the present guidelines, a
satisfactory criterion cannot be derived
at this time due to the Insufficency in the
available data for 1,2-dichloroethylene.

2,4-Dlchloropheno)

Freshwater Aquatic Life
  The available data for 2,4-
dichJorophenol indicate that acute and
chronic toxicity to freshwater aquatic
life occurs at concentrations as low as
2,020 and 365 fig/1, respectively, and
would occur at lower concentrations
among species that are more sensitive
that  those tested. Mortality to early life
stages of one species of fish occurs at
concentrations as low as 70 ug/l.

Saltwater Aquatic Life
  Only one test has been conducted
with saltwater organisms on 2,4-
dichJorophenol and no statement can be
made concerning acute or chronic
toxicity.

Human Health
  For comparison purposes, two
approaches were used to derive
criterion levels for 2,4-dichlorophenol.

-------
                   Federal Register  /  Vol. 45. No. 231  / Friday. November 28. 1980  / Notices
                                                                       70333
 Baaed on available toxicity data, for the
 protection of public health, the derived
 level is 3.09 mg/1. Using available
 organoleptic data, for controlling
 undesirable taste and odor quality of
 ambient water, the estimated level is 0.3
 pg/j. It should be recognized that
 organoleptic data as a basis for
 establishing a water quality criteria
 have limitations and have no
 demonstrated relationship to potential "
 adverse human "health effects,

 Dichloropropanes/DIchloropropenes

 Freshwater Aquatic Life

  The available data for '
 dichloropropanes Indicate that acute
 and chronic toxicity to freshwater
 aquatic life occurs at concentrations as
 low as 23,000 and 5,700 MS/I-
 respectively, and would occur at lower
 concentrations among species that are
 more sensitive than those tested.
  The available data for
 dichloropropenes indicate that acute
 and chronic toxicity to freshwater
 aquatic life occurs at concentrations as
 low as 6,060 and 244 ftg/1, respectively,
 and would  occur at lower
 concentrations among species that are
 more sensitive than those tested,

 Saltwater Aquatic Life

  The available data  for
 dichloropropanes indicate that acute
 and chronic toxicity to saltwater aquatic
 life occurs at concentrations as low as
 10,300 and 3,040 pg/L respectively, and
 would occur at lower concentrations
 among species that are more sensitive
 than those tested.  .
  The available data  for
 dlchJoropropenes Indicate that acute
 toxicity to saltwater aquatic life occurs
 at concentrations as low a as 780 fig/1,
 and would  occur at lower
 concentrations among spe'cies  that are
 more sensitive than those tested. No
 data are available concerning the
 chronic toxicity of dichloropropenes to
 sensitive saltwater aquatic life,

Human Health

  Using the present guidelines, a
 satisfactory criterion cannot be derived
 at this time due to the insufficiency in
 the available data for dichloropropanes.
  For the protection of human  health
 from the toxic properties of
 dichloropropenes ingested through
 water and contaminated aquatic
 organisms,  the ambient water criterion
 is determined to be 87 jig/I.
  For the protection of human  health
from the toxic properties of
dichloropropenes ingested through
 contaminated aquatic organisms alone.
the ambient water criterion is
determined to be 14.1 mg/1.
2,4-Dunethylphenol
Fresh water A qua tic Life
  The available data for 2,4-
dimethylphenol Indicate that acute
toxicity to freshwater aquatic life occurs
at concentrations as low as 2,120 jtg/1
and would occur at lower
concentrations among species .that are
more sensitive than those tested. No
data are available  concerning the
chronic toxicity of dlmethylpnenol to
sensitive freshwater aquatic life.

Saltwater Aquatic Life
  No saltwater organisms have been
tested with 2,4-dimethylphenol and no
statement can be made concerning acute
and chronic toxicity.

Human Health
  Sufficient data are not available for
2,4-dimethylphenol to derive a level
which would protect against the
potential toxicity of this compound.
Using available organoleptic data, for
controlling undersirable taste and  odor
quality of ambient  water, the estimated
level is 400 ug/1. It should be recognized
that organoleptic data as a basis for
establishing a water quality criteria
have limitations and have no
demonstrated relationship to potential
adverse human health effects.
2,4-Dlnitrotoluene

Freshwater Aquatic Life
  The available data for 2,4-
dinitrotoluene Indicate that acute and
chronic toxicity to  freshwater aquatic
life occurs at concentrations as low as
330 and 230 jtg/1, respectively, and
would occur at lower concentrations
among species that are more sensitive
than those tested.
Saltwater Aquatic  Life
  The available data for 2,4-
dinitrotoluenes indicate that acute
toxicity to saltwater aquatic life occurs
at concentrations as low as 590 fig/1 and
would occur at lower concentrations
among species that are more sensitive
than those tested. No data are available
concerning the chronic toxicity of 2,4-
dinitrotoluenes to sensitive saltwater
aquatic life but a decrease In algal cell
numbers occurs at  concentrations  as
low as 370 ftg/1.

Human Health
  For the  maximum protection of human
health from the potential carcinogenic
effects due to exposure of 2,4-
dinitrotoluene through ingestlon of
contaminated water and contaminated
 aquatic organisms, the ambient water
 concentration should be zero based on
 the non-threshold assumption for this
 chemical. However, zero level may not
 be attainable at the present time.
 Therefore, the levels which may result in
 incremental increase of cancer risk over
 the lifetime are estimated at  10~*. 10"«,
 and 10"'. The corresponding criteria are
 1.1 MS/I. O.ll.Mg/1. and 0.011 pg/1,
' respectively. If the above estimates are
 made for consumption of aquatic
 organisms only, excluding  consumption
 of water, the levels are 01 j*g/l,8.1 j*g/l,
 and 0.91 fig/1, respectively. Other
 concentrations representing different
 risk levels may be calculated by use of
 the Guidelines. The risk estimate range
 is presented for information purposes
 and does not represent an  Agency
 judgment on an "acceptable" risk level.

 1,2-Dlphenylhydrazlne

 freshwater Aquatic Life

   The available data for 1,2-
 diphenylhydrazlne indicate that acute
 toxicity to freshwater aquatic life occurs
 at concentrations as low as 270 fig/1 and
 would occur at lower concentrations
 among species that are more sensitive
 than those tested. No data are available
 concerning the chronic toxicity of 1,2-
 dlphenylhydraiine to sensitive
 freshwater aquatic life.

 Saltwater A qua tic Life

   No saltwater organisms  have been
 tested with 1,2-diphenylhydrazine and
 no statement can be made concerning
 acute and chronic toxicity.

 Human Health

   For the maximum protection of human
 health from the potential carcinogenic
 effects due to exposure of 1,2-
 diphenylhydrazine through ingestion of
 contaminated water and contaminated
 aquatic organisms, the ambient water
 concentration should be zero based on
 the non-threshold assumption for this
 chemical. However, zero level may not
 be attainable at the present time.
 Therefore, the levels which may result In
 incremental Increase of cancer risk over
 the lifetime are estimated at  10~*. 10~*.
 and 10"'. The corresponding  criteria are
 422 ng/1,42 ng/1, and 4 ng/1,
 respectively. If the  above estimates are
 made for consumption of aquatic
 organisms only, excluding  consumption
 of water, the levels are 5.8 ng/1. 0.56
 iug/l, and 0.058 jig/1, respectively.
 Other concentrations representing'
 different risk levels may be calculated
 by use of the Guidelines. The risk
 estimate range is presented for
 information purposes and does not

-------
 79334
Federal  Register / Vol. 45,  No. 231 /  Friday.  November 28. 1980  /  Notices
 represent an Agency judgment on an
 "acceptable" risk level.

 Endosulfan,
 Fresh water Aquatic Life
   For endosulfan the criterion to protect
 freshwater aquatic life as derived using
 the Guidelines is 0.056 Mg/1 as a 24-hour
 average and:the concentration should
 not exceed 0.22 jxg/1 at any time.

 Saltwater Aquatic Life
   For endosulfan the criterion to protect
 saltwater aquatic life as derived using,
 the Guidelines is 0.0087 fig/1 as a 24-
. hour average and the concentration
 should not exceed 0.034 jxg/1 at any
 time.
 Human Health
   For the protection of human health
 from the toxic properties of endosulfan
 ingested through water and
 contaminated aquatic organisms, the
 ambient water criterion is determined  to
 be 74 MS/1-
   For the protection of human health
 from the toxic properties of endosulfan
 ingested through contaminated aquatic
 organisms alone, the ambient water
 criterion is determined to be 159 jxg/1.

 Endiin
 Freshwater Aquatic Ufe
   For endrin the criterion to protect
 freshwater aquatic life as derived using
 the Guidelines is 0.0023 pg/ras a 24-
 hour average and the concentration
 should not exceed 6.18 jxg/1 at any tinie.

 Saltwater Aquatic Life           •
   For endrin the criterion to protect
 saltwater aquatic life as derived using
 the Guidelines is 0.0023 MS/I as a 24-
 hour average and the concentration
 should hot exceed 0.037 /ig/1 at any
 time.

 Human Health
   The ambient water quality criterion
 for endrin is recommended to be
 identical to the existing drinking water
 standard which  is 1 /ig/1. Analysis of the
 toxic effects data resulted In a
 calculated level  which is protective of
 human health against the ingestion of
 contaminated water and contaminated
 aquatic organisms. The calculated value
 is comparable to the present standard.
 For this reason a selective criterion
 based on exposure solely from
 consumption of 6.5 grams of aquatic
 organisms was not derived.
 Ethylbenzene
 Freshwater Aquatic Life
   The available data for ethylbenzene
 indicate that acute toxicity to freshwater
                     aquatic life occurs at concentrations as
                     low as 32,000 fig/1 and would occur at
                     lower concentrations among species
                     that are more sensitive than those
                     tested. No definitive data are available
                     concerning the chronic toxicity of
                     ethylbenzene to sensitive freshwater
                     aquatic life.

                     Saltwater Aquatic Life

                       The available data for ethylbenzene1
                     indicate that acute toxicity to saltwater
                     aquatic life occurs at concentrations as
                     low as 430 pg/1 and would occur at
                     lower concentrations among species
                     that are more sensitive than those
                     tested. No data are available concerning
                     the chronic toxicity of ethylbenzene to
                     sensitive saltwater aquatic life.

                     Human Health

                       For the protection of human health
                     from the toxic properties of
                     ethylbenzene ingested through water
                     and contaminated aquatic organisms,
                     the ambient water criterion is
                     determined to be 1.4 mg/1.
                       For the protection of human health
                     from the toxic properties of
                     ethylbenzene Ingested through
                     contaminated aquatic organisms alone,
                     the ambient water criterion is
                     determined to be 3.28 mg/1.  ,

                     Fluoranthene

                     Freshwater Aquatic Life

                       The available data for fluoranthene
                     indicate that acute toxicity to freshwater
                     aquatic life occurs at concentrations as
                     low as 3960 jxg/1 and would occur at
                     lower concentrations among species
                     that are more sensitive than those
                     tested. No data are available concerning
                     the chronic toxicity of fluoranthene to
                     sensitive freshwater aquatic life.
                     Saltwater Aquatic Life

                       The available data for fluoranthene
                     Indicate that acute and chronic toxicity
                     to saltwater aquatic life occur at
                     concentrations as low as 40 and 18 /ig/1,
                     respectively, and would occur at lower
                     concentrations among species that are
                     more sensitive than those tested.

                     Human Health

                       For the protection of human health
                     from the toxic properties of fluoranthene
                     ingested through water and
                     contaminated aquatic organisms, the
                     ambient water criterion is determined to
                     be42/tg/J.
                       For the protection of human health
                     from the toxic properties of fluoranthene
                     ingested through contaminated aquatic
                     organisms alone, the ambient water
                     criterion is determined to be 54 MS/'-
Haloethere

Freshwater Aquatic Life
  The available data for haloethers    •'
indicate that acute and chronic toxicity
to freshwater aquatic life occur at  •'•'••• •'",'
concentrations as low as 360 and 122
fig/I. respectively, and would occur at
lower concentrations among species
that are more sensitive than those
tested".'
Saltwater Aquatic Life
  No saltwater organisms have'been ..'•
tested with any haloether and no   .  .
statement can be made concerning acute
or chronic toxicity.

Human Health
  Using the present guidelines, a
satisfactory criterion cannot be derived
at this time due to the insufficiency in
the available data for haloethers.
Halomethanes
Freshwater Aquatic Life
  The available data for halomethanes
indicate that acute toxicity to freshwater
aquatic life occurs at concentrations as
low as 11.000 Mg/1 and would occur at
lower concentrations among species
that are more sensitive than those
tested. No data are available concerning
the chronic toxicity of halomethanes to :
sensitive freshwater aquatic life.
Saltwater Aquatic Life
  The available data for halomethanes
indicate that acute and chronic toxicity
to saltwater aquatic life occur at
concentrations as low as 12,000  and
6,400 Mg/1. respectively, and would
occur at lower concentrations among
species that are more sensitive than
those tested. A decrease in algal cell
numbers occurs at concentrations as
low as 11,500 fig/1.
Human Health
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of
chloromethane, bromomethane.
dichloromethane,
bromodichloromethane,
tribromomethane,
dichlorodifluoromethane,
trichlorofluoromethane, or combinations
of these chemicals through ingestion of
contaminated, water and contaminated
aquatic organisms, the ambient  water
concentration should be zero based on
the non-threshold assumption for this
chemical. However, zero level may not
be  attainable at the present time.
Therefore, the levels which may result in
incremental increase of cancer risk, over
the lifetimes are estimated at 10~5,10~6.
and 10"7. The corresponding criteria are

-------
                  Federal Register  /  Vol.  45. No. 231 / Friday. November 28. 1980 / Notices             79335
 1.9 fig/1, 0.19 ftg/1, and 0.019 fig/I,
 respectively. If the above estimates are
 made for consumption of aquatic
 organisms only, excluding consumption
 of water, the levels are 157 fig/1,15.7
 fig/1, and 1.57 ^g/1, respectively. Other
 concentrations representing different
 risk.levela may be calculated by use of
 the Guidelines. The risk estimate range
 is presented far information purposes
 and does not represent an Agency
 judgment on an  "acceptable" risk level.
 Heptachlor

 Freshwater Aquatic Life

  For heptachlor the criterion to protect
 freshwater aquatic life as derived using
 the Guidelines is 0.0038 jig/1 as a 24-
 hour average and the concentration
 should not exceed 0.52 fig/1 at any time.

 Saltwater Aquatic Life

  For heptachlor the criterion to protect
 saltwater aquatic life as derived using
 the Guidelines is 0.0036 f»g/l aa a 24-
 hour average and the concentration
 should not exceed 04)53 fig/1 at any
 time.
 Human Health

  For the maximum protection of human
 health from the potential carcinogenic
 effects due to exposure of heptachlor
 through ingestion of contaminated water
 and contaminated aquatic organisms,
 the ambient water concentration should
 be zero based on the non-threshold
 assumption for this chemical. However,
 zero level may not be attainable at the
 present time. Therefore, the levels which
 may result In Incremental Increase of
 cancer risk, over the lifetimes are
 estimated at 10"», 10'«, and 10"f. The
corresponding criteria are 2.78 ng/1, .28
ng/1, and .028 ng/1, respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the
levels are 2.85 ng/1, ,28 ng/1, and ,028
 ng/1, respectively. Other
concentrations representing different
risk levels may be calculated by use of
the Guidelines. The risk estimate range
is presented for information purposes
and does not represent an Agency
judgment on an "acceptable" risk level.
Hexachlorobutadlene

 Freshwater Aquatic Life

  The available data for
 hexachlorobutadiene indicate  that acute
and chronic/toxicity to freshwater
aquatic life occur at concentrations as
 low as 90 and 9.3 fig/1, respectively, and
 would occur at lower concentrations
 among species that are more sensitive
 •-ban those tested.
Saltwater Aquatic Life
  The available data for
hexachlorobutadiene indicate that acute
toxicity to saltwater aquatic life occurs
at concentrations as low as 32 fig/1 and
would occur at lower concentrations
among species that are more sensitive
that those tested. No data are available
concerning the chronic toxicity of
hexachlorobutadiene to sensitive
saltwater aquatic life

Human Health
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of
hexachlorobutadiene through ingestion
of contaminated water and
contaminated aquatic organisms, the
ambient water concentration should be
zero based on the non-threshold
assumption for this chemical. However,
zero level may not be attainable at the
present time. Therefore, the levels which
may result in incremental increase of
cancer risk, over the lifetimes are
estimated at 10"5,10"8, and 10" \ The
corresponding criteria are 4.47 fig/1,0.4S
jig/1, and 0.045 fig/1, respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the
levels are 900 fig/1,50 fig/1, and 6 fig/1
respectively. Other concentrations
representing different risk levels may be
calculated by use of the Guidelines. The
risk estimate range is presented for
information purposes and does not
represent an Agency Judgment on an
"acceptable" risk level.
Hexachlorocydohexane
Undone
Freshwater Aquatic Life
  For Lindane the criterion to protect
freshwater aquatic life aa derived using
the Guidelines la 0.080 fig/1 as a 24-hour
average and the concentration should
not exceed Z.O.fig/l at any time.
Saltwater Aquatic Life
  For saltwater aquatic life the
concentration of lindane should not
exceed 0.16 fig/1 at any time. No data
are available concerning the chronic
toxicity of lindane to sensitive saltwater
aquatic life.
BHC
Freshwater Aquatic Life
  The available date for a mixture of
isomers of BHC indicate that acute
toxicity to freshwater aquatic life occurs
at concentrations as low as 100 fig/1 and
would occur at lower concentrations
among species that are more sensitive
than those tested. No data are available
concerning the chrorJc tox:ci;y of a
mixture of isomers of BHC to sensitive
freshwater aquatic life.
Saltwater Aquatic Life

  The available date for a mixture of
isomers of BHC indicate that acute
toxicity to saltwater aquatic life occurs
at concentrations as low as 0.34 fig/1
and would occur at lower
concentrations among species that are
more sensitive than those tested. No
data are available concerning the
chronic toxicity of a mixture of isomers
of BHC to sensitive saltwater aquatic
life.
Human Health

  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of alpha-HCH
through ingestion of contaminated water
and contaminated aquatic organisms,
the ambient water concentration should
be zero based on the non-threshold
assumption for this chemical. However,
zero level may not be attainable at the
present time. Therefore, the levels  which
may result in  incremental  increase of
cancer risk, over the lifetimes are
estimated at 10"8.10~«, and 10"'. The
corresponding criteria are 92 ng/1,9.2
ng/1, and .92 ng/1, respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the
levels are 310 ng/1, 31.0 ng/1, and 3.1
ng/1 respectively. Other concentrations
representing different risk levels may be
calculated by use of the Guidelines. The
risk estimate range Is presented for
Information purposes and does not
represent an Agency judgment on an
"acceptable"  risk level.
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of beta-HCH
through ingestion cf contaminated water
and contaminated aquatic organisms,
the ambient water concentration should
be zero based on the non-threshold
assumption for tola chemical. However,
zero level may not be attainable at the
present time. Therefore, the levels which
may result In  incremental  increase of
cancer risk, over the lifetimes are
estimated at lO"8.1CT«, and 10'*. The
corresponding criteria are 163 ng/L 16.3
ng/l,-and 1.63 ng/1, respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the
levels are 547 ng/1,54.7 ng/L and 5.47
ng/1, respectively. Other concentrations
representing different risk levels may be
calculated by use of the Guidelines. The
risk estimate range is presented for
information purposes and does not

-------
 79336
                   Federal Register / Vol. 45. No.  231 / Friday. November 28. 1980 / Notices
 represent an Agency Judgment on an
 "acceptable" risk level.
   For the maximum protection of human
 health from the potential carcinogenic
 effects due to exposure of tech-HCH
 through ingestion of contaminated water
 and contaminated aquatic organisms,
 the ambient water concentration should
 be zero based on the non-threshold
 assumption for this chemical. However.
 zero level may not be attainable at the
 present time. Therefore, the levels which
 may result in Incremental increase of
 cancer risk, over the lifetimes are
 estimated at 10~s. ID"6, and 10"T. The
 corresponding criteria are 123 ng/l. 12.3
 ng/1. and T.23 ng/l, respectively. If the
 above estimates are made for
 consumption of aquatic organisms only,
 excluding consumption of water, the
 levels are 414 ng/l, 41.4 ng/l, and 4.14
 ng/l, respectively.  Other concentrations
 representing different risk levels may be
 calculated by use of the Guidelines.  The
 risk estimate range is  presented for
 information purposes  and does not
 represent an Agency judgment on an
 "acceptable" risk level.
  For the maximum protection of human
 health from the potential carcinogenic
 effects due to exposure of gamma-HCH
 through Ingestion of contaminated water
 and contaminated  aquatic organisms,
 the ambient water concentrations
 should be zero based on the non-
 threshold assumption  for this chemical.
 However, zero level may not be
 attainable at the present time.
 Therefore, the levels which may result in
 incremental increase of cancer risk over
 the lifetime are estimated at ICT", 10~6,
 and 1CT7. The corresponding criteria  are
186 ng/l 18.6 ng/l,  and 1.86 ng/l
respectively. If the above estimates are
made for consumption of aquatic
organisms only, excluding consumption
of water, the levels are 625 ng/l, 62.5
 ng/l, 6.25 ng/l, respectively. Other
concentrations representing different
risk levels may be calculated by use  of
the Guidelines. The risk estimate range
is presented for information purposes
and does not represent an Agency
judgment on an "acceptable" risk level.
  Using the present guidelines, a
satisfactory criterion cannot be derived
at this time due to the  insufficiency in
the available data for  delta-HCH.
  Using the present guidelines, a
satisfactory criterion cannot be derived
at this time due to the insufficiency in
the available data for  epsilon-HCH.

Hexachlorocyclopentadlene

Freshwater Aquatic Life
  The available data for
hexachlorocyf lopentadiene indicate that
acute and chronic toxicity to freshwater
 aquatic life occurs at concentrations as
 low as 7.0 and 5.2 jig/1, respectively, and
 would occur at lower concentrations
 among species that are more sensitive
 than those tested.
 Saltwater Aquatic Life
  The available data to
 hexachlorocyclopentadiene indicate that
 acute toxicity to saltwater aquatic life
 occurs at concentrations as low as 7.0.
 p.g/1 and would occur at lower
 concentrations among species that are
 more sensitive than those tested. No
 data are available concerning the
 chronic toxicity of
 hexachlorocyclopentadiene to sensitive
 saltwater aquatic life.
Human Health
  For comparison purposes, two
 approaches were used to derive
 criterion levels for
 hexachlorocyclopentadiene. Based on
 available toxicity data, for the
 protection of public health, the derived
 level is 206 jig/1. Using available
 organoleptic data, for controlling
undesirable taste and odor quality of
 ambient water, the estimated level is 1.0
fig/1. It should be recognized that
organoleptic data as a basis for
establishing a water quality criterion
have limitations and have no
demonstrated relationship to potential
adverse human health effects.
Isophorone
Freshwater Aquatic Life
  The available data for isophorone
indicate that acute toxicity to freshwater
aquatic life ocurs at concentrations as
low as 117,000 M8/1 and would occur at
lower concentrations among species
that are more sensitive than those
tested. No data are available concerning-
the chronic toxicity of isophorone to
sensltive'freshwater aquatic life.
Saltwater Aquatic Life  .
  The available data for isophorone
indicate that acute toxicity to saltwater
aquatic life occurs at concentrations as
low as 12,900 jig/1 and would occur at
lower concentrations among species
that are more sensitive than those
tested. No data are available concerning
the chronic toxicity of isophorone to
sensitive saltwater aquatic life.

Human Health
  For the protection of human health
from the toxic properties of isophorone
ingested through water and
contaminated aquatic organisms, the
ambient water criterion is determined to
be 5.2 mg/1.
  For the protection of human health
from the toxic properties of isophorone
ingested through contaminated aquatic
organisms alone, the ambient water
criterion is determined to be 520 mg/L
Lead
                          *
Freshwater Aquatic Life
  For total recoverable lead the
criterion (in jig/1) to protect freshwater
aquatic life as derived using the
Guidelines is the numerical value given
by e(2.35[ln(hardness)]-9.48) as a 24-
hour average and the concentration (in
fig/1) should not exceed the numerical
value given by e(1.22[ln(hardness)]-0.47J
at any time. For example, at hardnesses
of 50,100, and 200 mg/1 as CaCO, the
criteria are 0.75,3.8. and 20 fig/1.
respectively,  as 24-hour averages, and
the concentrations should not exceed 74,
170. and 400 fig/I, respectively, at any
time.

Saltwater Aquatic Life
  The available data for total
recoverable lead indicate that acute and
chronic toxicity to saltwater aquatic life
occur at concentrations as low as 666
and 25 fig/l, respectively, and would
occur at lower concentrations among
species that are more sensitive than
those tested.

Human Health
  The ambient water quality criterion
for lead is recommended to be identical
to the existing drinking water standard
which is 50 fig/1. Analysis of the toxic
effects data resulted in a calculated
level which is protective to human
health against the ingestion of
contaminated water and contaminated
aquatic organisms. The calculated value
is comparable to the present standard.
For this reason a selective criterion
based on exposure solely from
consumption  of 6.5 grams of aquatic
organisms was not derived.

Mercury

Freshwater Aquatic Life
  For total recoverable mercury the
criterion to protect freshwater aquatic
life as derived using the Guidelines is
0.00057 fig/1 as a 24-hour average and
the concentration should not exceed
0.0017 fig/1 at any time.

Saltwater Aquatic Life
  For total recoverable mercury the
criterion to protect saltwater aquatic life
as derived using the Guidelines is 0.025
fig/I as a 24-hour average and the
concentration should not exceed 3.7 fig/I
at any time.

Human Health
  For the protection of human health
from the toxic properties of mercury

-------
                  Federal Register  /  Vol. 45. No. 231 / Friday. November  28. 1980 / Notices
                                                                       79337
ingested through water and
contaminated aquatic organisms, the
ambient water criterion is determined to
be 144 ng/1.
  For the protection of human health
from the toxic.properties of mercury
ingested through contaminated aquatic
organisms alone, the ambient water
criterion is determined to be 146 ng/1.
  Note.—These values Include the
consumption of freshwater, estuarine. and
marine species.

Naphthalene

Freshwater Aquatic Life

  The available data to naphthalene
indicate that acute and chronic toxicity
to freshwater aquatic life occur at
concentrations as low as 2.300 and 620
•j.g/1 respectively, and would occur at
lower concentrations among species
that are more sensitive than those
tested.

Saltwater Aquatic Life

  The available data for naphthalene
indicate that acute toxicity to saltwater
aquatic life occurs at concentrations as
low as 2,350 Mg/1 and would occur at
lower concentrations among species
 hat are more sensitive than those
 ested. No-data are available concerning
 he chronic toxicity of naphthalene to
.snsltive saltwater aquatic life.

 iuman Health

  Using the present guidelines, a
 atisfactory criterion cannot be derived
 t this time due to the insufficiency in
 .ie available data for naphthalene.

 i'ickel

 'reshwaterAquatic Life

  For total recoverable nickel the
 riterion (in Mg/1) to protect freshwater
 quatic life as derived using the
  tiidelines is the numerical value given
 y e(0.76 [In (hardness)] +1.06) as a 24-
 aur average and the concentration (in
 3/1) should not exceed the numerical
 ilue given by e(0.76[ln (hardness)] +
 02) at any time. For example, at
 irdnesses of SO, 100, and 200 mg/1 as
 aCO, the criteria are SO, 06, and 160
 ;/']. respectively, as 24-hour averages,
  id the concentrations should not
  ceed 1.100.1,800, and 3,100 jig/1,
  apectively, at any time.

  i(water Aquatic Life

  For total recoverable nickel the
  terion to protect saltwater aquatic life
  derived using the  Guidelines is 7.1
  /I as a 24-hour average and the
  acentration should not exceed 140 /ig/
  t any time.
Human Health
  For the protection of human health
from the toxic properties of nickel
ingested through water and
contaminated aquatic organisms, the
ambient water criterion is determined to
be 13.4 Mg/1-
  For the protection of human health
from the toxic properties of nickel
ingested through contaminated aquatic
organisms alone, the ambient water
criterion is determined to be 100 /ig/1.

Nitrobenzene
Freshwater Aquatic Life
  The available data for nitrobenzene
indicate that acute toxicity to freshwater
aquatic life occurs at concentrations as
low as 27,000 Mg/1 and would occur at
lower concentrations among species
that are more sensitive than those
tested. No definitive data are available
concerning the chronic toxicity of
nitrobenzene to sensitive freshwater
aquatic life.

Saltwater A quatic Life
  The available data for nitrobenzene
indicate that acute toxicity to saltwater
aquatic life occurs at concentrations as
low as 0,680 Mg/1 and would occur at
lower concentrations among species
that are more sensitive than those
tested. No data are available concerning
the chronic toxicity of nitrobenzene to
sensitive saltwater aquatic life.
Human Health
  For comparison purposes, two
approaches were used to derive
criterion levels for nitrobenzene. Based
on available toxicity data, for the
protection of public health, the derived
level is 19.8 mg/1. Using  available
organoleptic data, for controlling
undesirable taste and odor quality of
ambient water, the estimated level is 30
^g/1. It should be recognized that
organoleptic data as a basis for
establishing a water quality criteria
have limitations and have no
demonstrated relationship to potential
adverse human health effects.
Nltrophenols
Freshwater Aquatic Life
  The available data for nitrophenols
indicate that  acute toxicity to freshwater
aquatic life occurs at concentrations as
low as 230 /ig/1 and would occur at
lower concentrations among species
that are more sensitive than those
tested. No data are available concerning
the chronic toxicity of nitrophenols to
sensitive freshwater aquatic life but
toxicity to one species of algae occurs at
concentrations as low as 150 Mg/1-
 Saltwater Aquatic Life
   The available data for nitrophenols
 indicate that acute toxicity to saltwater
 aquatic life occurs at concentrations as
 low as 4.850 MS/' &na would occur at
 lower concentrations among species
 that are more sensitive than those
 tested. No data are available concerning
 the chronic toxicity of nitrophenols to
 sensitive saltwater aquatic life.

 Human Health
   For the protection of human health
 from the toxic properties of 2,4-dinitro-o-
 cresol ingested through water and
 contaminated aquatic organisms, the
 ambient water criterion is determined  to
 be 13.4 Mg/1.
   For the protection of human health
 from the toxic properties of 2,4-dinitro-o-
 cresol ingested through contaminated
 aquatic organisms alone, the ambient
 water criterion is determined to be 765
   For the protection of human health
 from the toxic properties of
 dinitrophenol ingested through water
 and contaminated aquatic organisms.
 the ambient water criterion is
 determined to be 70 Mg/1.
   For the protection of human health
 from the toxic properties of
 dinitrophenol ingested through
 contaminated aquatic organisms alone.
 the ambient water criterion is
 determined to be 14.3 mg/1.
   Using the present guidelines, a
 satisfactory criterion cannot be derived
 at this time due to the insufficiency in
 the available data for mononitrophenol.
   Using the present guidelines, a
 satisfactory criterion cannot be derived
 at this time due to the insufficiency in
 the available data for tri-nitrophenol.

 Nltrosamines
 Fresh water Aquatic Life
   The available data for nitrosamines
 indicate that acute toxicity to freshwater
 aquatic life occurs at concentrations as
• low as 5,850 Mg/1 and would occur at
 lower concentrations among species
 .that are more sensitive than those
 tested. No data are available concerning
 the chronic toxicity of nitrosamines to
 sensitive freshwater aquatic life.

 Saltwater Aquatic Life
   The available data for nitrosamines
 indicate that acute toxicity to saltwater'
 aquatic life occurs at concentrations as
 low as 3.300,000 Mg/1 and would occur at
 lower concentrations among species '
 that are more sensitive than those
 tested. No data are available concerning
 the chronic toxicity of nitrosamines to
 sensitive saltwater aquatic  life.

-------
 79338
Federal Register / Vol. 45. No. 231 /  Friday.  November 28. 1980 / Notices
 Human Health
  For the maximum protection of human
 health from the potential carcinogenic
 effects due to exposure of n-
 nitrosodlmethylamine through ingestion
 of contaminated water and
 contaminated aquatic organisms, the
 ambient water concentration should be
 zero based on the non-threshold
 assumption for this chemical.-However,
 zero level may not be attainable at the
 present time. Therefore, the levels which
 may result in incremental increase of
 cancer risk, over the lifetimes are
 estimated at I0~», W8, and 10~f. The
 corresponding criteria are 14 ng/1,1,4
 ng/1, and .14 ng/1, respectively. If the
 above estimates are made for
 consumption of aquatic organisms only,
 excluding consumption of water, the
 levels are 180,000 ng/1,16,000 ng/1, and
 1,600 ng/1, respectively. Other
 concentrations representing different
 risk levels may be calculated by use of
 the Guidelines. The risk estimate range
 is presented for Information purposes
 and does not represent an Agency
 judgment on an "acceptable" risk level.
  For the maximum protection of human
 health from the potential carcinogenic
 effects due to exposure of n-
 nitrosodiethylamine through ingestion of
 contaminated water and contaminated
 aquatic organisms, the ambient water   ~
 concentration should be zero based on
 the non-threshold assumption for this
 chemical However, zero level may not
 be~attainable at the present time.
 Therefore, the levels which may result in
 incremental increase of cancer risk, over
 the lifetimes are  estimated at 10~*. 10~c,
and 10"'. The corresponding criteria are
8 ng/1,0.8 ng/1, and 0.08 ng/1,
respectively. If the above estimates are
made for consumption of aquatic
 organisms only, excluding consumption
of water, the levels are 12,400 ng/1, 1,240
ng/1, and 124 ng/L respectively. Other
concentrations representing different •
risk levels may be calculated by use of
the Guidelines. The risk estimate range
Is presented for information purposes
and doe's not represent an Agency
judgment on an "acceptable" risk level.
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure in n-nitrosodi-n-
butylamine through ingestion of
contaminated water and contaminated
aquatic organisms, the ambient water
concentration should be zero based on
 the non-threshold assumption for this
chemical. However, zero level may not
 be attainable at the present time.
Therefore, the levels which may result in
incremental  increase of cancer risk, over
the lifetimes are estimated at 10~s, 10~*.
and 10"'. The corresponding criteria are
                     64 ng/16.4 ng/1 and .064 ng/1,
                     respectively. If the above estimates are
                     made for consumption of aquatic
                     organisms only, excluding consumption
                     of water, the levels are 5,868 ng/1, 587
                     ng/1, and 58.7 ng/1, respectively. Other
                     concentrations representing different
                     risk levels may be calculated by use of
                     the Guidelines. The risk estimate range
                     is presented for information purposes
                     and does not represent an Agency
                     judgment on an "acceptable" risk level.
                       For the maximum protection of human
                     health from the potential  carcinogenic
                     effects due to exposure In n-
                     nitrosodiphenylamine through Ingestion
                     of contaminated water and
                     contaminated aquatic organisms, the
                     ambient water concentration should be
                    * zero based on the non-threshold
                     assumption for this chemical. However,
                     zero level may not be attainable at the
                     present time. Therefore, the levels which
                     may result in incremental increase of
                     cancer risk, over the lifetimes are
                     estimated at 10""°. 10"*, and 10"I The
                     corresponding criteria are 48,000 ng/1
                     4,900 ng/1 and 490 ng/I, respectively. If
                     the above estimates are made for
                     consumption of aquatic organisms only,
                     excluding consumption of water,  the
                     levels are 161,000 ng/1,16.100 ng/1,  and
                     1,610 ng/1, respectively. Other
                     concentrations  representing different
                     risk levels may be calculated by use of
                     the Guidelines. The risk estimate range
                     is presented for information purposes
                     and does not represent an Agency
                     judgment on an "acceptable" risk level.
                       For the maximum protection of human
                     health from the potential  carcinogenic
                     effects due to exposure in n-
                     nitrosopyrrolidine through ingestion of
                     contaminated water and contaminated
                     aquatic organisms, the ambient water
                     concentration should be zero based on
                     the non-threshold assumption for this
                     chemical. However, zero  level may not
                     be attainable at the present time.
                     Therefore, the levels which may result in
                     incremental increase of cancer risk, over
                     the lifetimes are estimated at 10"*, 10"*
                     and 10"7. The corresponding criteria are
                     160 ng/116.0 ng/1 and 1.60 ng/1.
                     respectively. If the above estimates are
                     made for consumption of aquatic
                     organisms only, excluding consumption
                     of water, the levels are 919.000 ng/1,
                     91.900 ng/1,  and 9,190 ng/1. respectively.
                     Other concentrations representing
                     different risk levels may be calculated
                     by use of the Guidelines.  The risk
                     estimate range  is presented for
                     information purposes and does not
                     represent an Agency judgment on an
                     "acceptable" risk level.
Pentachlorophenol

Fresh water Aquatic Life
  The available data for     •
pentachlorophenol indicate that acute
and chronic toxicity to freshwater
aquatic life occur at concentrations as
low as 55 and 3.2 pg/1, respectively, and
would occur at lower concentrations
among species that are more sensitive
than those tested.
Saltwater Aquatic Life
  The available data for
pentachlorophenol Indicate that acute
and chronic toxicity to saltwater aquatic
life occur at concentrations' as low as 53
and 34 MS/1- respectively, and would
occur at lower concentrations among
species that are more sensitive  than
those tested.
Human Health
  For comparison purposes, two
approaches were used to derive
criterion levels for pentachlorophenol.
Based on available toxicity data, for the
protection of public health, the derived
level is 1.01 ing/1. Using available
organoleptic data, for controlling
undesirable taste and odor quality of
ambient water, the estimated level is 30
Hg/l.'It should be recognized that
organoleptic data as a basis for
establishing a water quality criterion
have limitations  and have no
demonstrated relationship to potential
adverse human health effects.

Phenol
Fresh water A qua tic Life
  The available data for phenol indicate
that  acute and chronic toxicity  to
freshwater aquatic life occur at
concentrations as low as 10.200 and
2,580 jtg/1, respectively, and would
occur at lower concentrations among
species that are more sensitive  than
those tested,

Saltwater Aguatic Life
  The available data for phenol indicate
that  acute toxicity to saltwater  aquatic
life occurs at concentrations as low as
5,800 ug/1 and would occur at lower
concentrations among species that are
more sensitive than those tested. No
data are available concerning the
chronic toxicity of phenol to sensitive
saltwater aquatic life.
Human Health
  For comparison purposes, two
approaches were used to derive
criterion levels for phenol. Based on
available toxicity data, for the
protection of public health, the  derived
level is 3.5 rng/1. Using available
organoleptic data, for controlling

-------
                   Federal Register /  Vol.  45,  No. 231  /  Friday, November 28. 1980 /  Notices
                                                                       79339
undesirable taste and odor quality of
ambient water, the estimated level Is 0.3
mg/1. It should be recognized that
organoleptic data as a basis for
establishing a water quality criterion
have limitations and have no
demonstrated relationship to potential
adverse human health effects.

Phthalate Esters

Freshwater Aquatic Life
  The available data for phthalate
esters indicate that acute and chronic
toxidty to freshwater aquatic life occur
at concentrations as low as 940 and 3
fig/1, respectively, and would occur at
lower concentrations among species
that are more sensitive than those
tested.

Saltwater Aquatic Life
  The available data for phthalate
esters indicate that acute toxicity to
saltwater aquatic life occurs at
concentrations as low as 2944 fig/1 and
would occur at lower concentrations
among species that are more sensitive
than those tested. No data are available
concerning the chronic toxicity of
phthalate esters to sensitive saltwater
aquatic life but toxicity to one species of
algae occurs at concentrations as low as
Human Health
  For the protection of human health
from the toxic properties of dimethyl-
phthalate ingested through water and
contaminated aquatic organisms, the
ambient water criterion is determined to
be 313 mg/1.
  For the protection of human health
from the toxic properties of dimethyl-
phthalate ingested through
contaminated aquatic organisms alone,
the ambient water criterion is
determined to be 2.9 g/1.
  For the protection of human health
from the toxic properties of diethyl-
phthalate ingested through water and
contaminated aquatic organisms, the
ambient water criterion is determined to
be 350 mg/1.
  For the protection of human-health
from the toxic properties of diethyl-
phthalate ingested through
contaminated aquatic organisms alone,
the ambient water criterion is
determined to be 1.8 g/1.
  For the protection of human health
from the toxic properties of dibutyl-
phthalate ingested through water and
contaminated aquatic organisms, the
ambient water criterion is determined to
be 34 mg/1.
  For the protection of human health
from the toxic properties of dibutyl-
phthalate ingested through
contaminated aquatic organisms alone,
the ambient water criterion is
determined to be 154 mg/1.
  For the protection of human health
from the toxic properties of di-2-
ethylhexyl-phthalate ingested through
water and contaminated aquatic
organisms, the ambient water criterion
is determined to be 15 mg/L
  For the protection of human health
from the toxic properties of di-2-
ethylhexyl-phthalate ingested through
contaminated aquatic organisms alone,
the ambient water criterion is
determined to be 50 mg/1.

Polychlorinated Biphenyls

Fresh water A qua tic Life

  For polychlorinated biphenyls the
criterion to protect freshwater aquatic
life as derived using the Guidelines is
0.014 fig/1 as a 24-hour average. The
available data indicate that acute
toxicity to freshwater aquatic life
probably will only occur at
concentrations above 2.0 fig/1 and that
the 24-hour average should provide
adequate protection against acute
toxicity.

Saltwater Aquatic Live

  For po'ychlorinated biphenyls the
criterion to protect saltwater aquatic life
as derived using the Guidelines is 0.030
jjg/1 as a 24-hour average. The available
data indicate that acute toxicity to
saltwater aquatic life probably will only
occur at concentrations above 10 fig/1
and that the 24-hour average should
provide adequate protection against
acute toxicity.
Human Health

  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure -of PCBs through
ingestion of contaminated water and
contaminated aquatic organisms, the
ambient water concentration should be
zero based on the non-threshold
assumption for this chemical. However,
zero level may not be attainable at the
present time. Therefore, the levels which
may result in incremental increase of
cancer risk over the lifetime are
estimated at 10"', 10~*. and  10"'. The
corresponding criteria are .79 ng/1,0.79
ng/1, and .0079 ng/1, respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the
levels are .79 ng/1, .079 ng/1, and .0079
ng/1. respectively. Other concentrations
representing different risk levels may be
calculated by use of the Guidelines. The
risk estimate range is presented for
information purposes and does not
represent an Agency judgment on an
"acceptable" risk level.
Polynuclear Aromatic Hydrocarbons
(PAHs)
Freshwater Aquatic Life
  The limited freshwater data base
available for polynuclear aromatic
hydrocarbons, mostly from short-term
bloconcentration studies with two
compounds, does not permit a statement
concerning acute or chronic toxicity.

Saltwater Aquatic Life
  The available data for polynuclear
aromatic hydrocarbons indicate that
acute toxicity to saltwater aquatic life
occurs at concentrations as low as 300
ug/1 and would occur at lower
concentrations among species that are
more sensitive than those tested. No
data are available concerning the
chronic toxicity of polynuclear aromatic
hydrocarbons to sensitive saltwater
aquatic life.
Human Health
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of PAHs through
ingestion of contaminated water and
contaminated aquatic organisms, the
ambient water concentration should be
zero based on the non-threshold
assumption for this chemical. However,
zero level may not be attainable at the
present time. Therefore, the levels which
may result in incremental increase of
cancer risk over the lifetime are
estimated at 10"s. 10'6. and 10"'. The
corresponding criteria are 28 ng/1, 2.8
ng/1, and .28 ng/1, respectively. If the
above estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the
levels are 311 ng/1. 31.1 ng/1, and 3.11
ng/1, respectively. Other concentrations
representing different risk levels may be
calculated by use of the Guidelines. The
risk estimate range is presented for
information purposes and does not
represent an Agency judgment on an
"acceptable" risk level.
Selenium
Freshwater Aquatic Life
  For total recoverable inorganic
selenite the criterion to protect
freshwater aquatic life as derived using
the Guidelines is 35 fig/1 as a 24-hour
average and the concentration should
not exceed 260 pg/1 at any time.
  The available data for inorganic
selenate indicate  that acute toxicity to
freshwater aquatic life occurs at
concentrations as low as 760 pig/1 and
would occur at lower concentrations
among species that are more sensitive

-------
79340
                   Federal  Register / Vol. 45. No. 231  / Friday. November 2B.  1980 / Notices
than those tested. No data are available
concerning the chronic toxicity of
inorganic selenate to sensitive
freshwater aquatic life.
Saltwater Aquatic Life
  For total recoverable inorganic
selenlte the criterion to protect saltwater
aquatic life as derived using the
Guidelines is 54 Mg/1 *»» « 24-hour
average and the concentration should
not exceed 410 jtg/1 at any time.   .
  No data are available concerning the
toxicity of inorganic selenate to
saltwater aquatic life.

Human Health
  The ambient water quality criterion
for selenium is recommended to be
identical to the existing drinking .water
standard which is 10 fig/1. Analysis of
the toxic effects data resulted in a
calculated level which is protective of
human health against the Ingestion of
contaminated water and contaminated
aquatic organisms. The calculated value
Is comparable to the present standard.
For this reason a selective criterion
based on exposure solely from
consumption of 6.5 grams of aquatic
organisms was not derived.
Silver
Freshwater Aquatic Life
  For freshwater aquatic life the
concentration (in jig/1) of total
recoverable silver should not exceed the
numerical value given by "e[1.72(ln
(hardness}-6.52}]" at any time. For
example, at hardnesses of 50,100,200
rng/1 as CaCOi the concentration of
total recoverable silver should not
exceed 1.2,4.1, and 13 jig/1, respectively,
at any time. The available data indicate
that chronic toxicity to freshwater
aquatic life may occur at concentrations
as low as 0.12 fig/1.

Saltwater Aquatic Life
  For saltwater aquatic life the
concentration of total recoverable silver
should not exceed 2.3 fig/1 at any time.
No data are available concerning the
chronic toxicity of silver to sensitive
saltwater aquatic life.
Human Health
  The ambient water quality criterion
for silver is recommended to be
identical to the existing drinking water
standard which is SO jtg/1. Analysis of
the toxic effects data resulted in a
calculated level which is protective of
human health against the ingeation of
contaminated water and contaminated
aqua lie organisms. The calculated value
is comparable to the present standard.
For this reason a selective criterion
based on exposure solely from
consumption of 6.5 grams of aquatic
organisms was not derived,
Tetrachloroethylene
Freshwater Aquatic Life
  The available data for
tetrachloroethylene Indicate that acute
and chronic  toxicity to freshwater
aquatic life occur at concentrations as
low as 5,280 and 840 fig/1, respectively.
and would occur at lower
concentrations among species that are
more sensitive than those tested.
Saltwater Aquatic Life
  The available data for
tetrachloroethylene indicate that acute
and chronic  toxicity to saltwater aquatic
life.occur at  concentrations low as
10,200 and 450 fig/1, respectively, and
would occur at lower concentrations .
among species that are more sensitive
than those tested.
Human Health
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of
tetrachloroethylene through ingestion of
contaminated water and contaminated
aquatic organisms, the ambient water
concentration should be zero based on
the non-threshold assumption  for this
chemical. However, zero level may not
be attainable at the present time.
Therefore, the levels which may result in
incremental  increase of cancer risk over
the lifetime ore estimated at 10~*. 10~*,
and 10~T. The corresponding criteria are
8 fig/1, .8 jig/1, and .08 fig/1, respectively.
If the above  estimates are made for
consumption of aquatic organisms only,
excluding consumption of water, the
levels are 88.5 fig/1,8.85 fig/1, and ,88
fig/1, respectively. Other concentrations
representing different risk levels may be
calculated by use of the Guidelines. The
risk estimate range is presented for
information purposes and does not
represent an Agency judgment on an
"acceptable" risk level.
Thallium

Freshwater Aquatic Life
  The available data for thallium
indicate that acute and chronic toxicity
to freshwater aquatic life occur at
concentrations as low as 1,400 and 40
ug/1, respectively, and would occur at
lower concentrations among species
that are more sensitive than those •
tested. Toxicity to one species of fish
occurs at concentrations as low as 20
fig/1 after 2,600 hours of exposure.

Saltwater Aquatic Life
  The available data for thallium
indicate that acute toxicitv to  saltwater
aquatic life occurs at concentrations as
low as 2,130 fig/1 and would occur at
lower concentrations among species
that are more sensitive than those
tested. No data are available Concerning
the chronic toxicity of thallium to
sensitive saltwater aquatic life.
Human Health
  For the protection of human health
from the  toxic properties of thallium
ingested  through water and
contaminated aquatic organisms, the
ambient water criterion is determined to
be 13 fig/1-
  For the protection of human health
from the  toxic properties of thallium
ingested  through contaminated aquatic
organisms alone, the ambient water
criterion  is determined to be 4B fig/1.

Toluene
Freshwater Aquatic Life
  The available data for toluene
indicate that acute toxicity to freshwater
aquatic life occurs at concentrations as
low as 17.500 fig/1 and would occur at
lower concentrations among species
that are more sensitive than those
tested. No data are available concerning
the chronic toxicity of toluene to
sensitive freshwater aquatic life.
Saltwater Aquatic Life
  The available data for toluene
indicate that acute and chronic toxicity
to saltwater aquatic life occur at
concentrations as low as 6,300 and 5,000
fig/1, respectively, and would occur at
lower concentrations among species
that are more sensitive than those
tested.

Human Health
  For the protection of human health
from the  toxic properties of toluene
ingested  through water and
contaminated aquatic organisms, the
ambient water criterion is determined to
be 14.3 mg/1.
  For the protection of human health
from the  toxic properties of toluene
ingested  through contaminated aquatic
organisms alone, the ambient water
criterion  is determined to be 424 mg/1.
Toxaphene

Freshwater Aquatic Life
  For toxaphene the criterion to protect
freshwater aquatic life as derived using
the Guidelines is 0.013 fig/1 as a 24-hour
average and the concentration should
not exceed 1.6 ftg/1 at any time.
Saltwater A guatic Life
  For saltwater aquatic life the
concentration of toxaphene should not
exceed 0.070 ftg/1 at any time. No data

-------
                  Federal Register / Vol. 45. No. 231 / Friday, November 28. i960 / Notices
                                                                      79341
 are available concerning the chronic
 tcxicity of toxaphone to sensitive
 saltwater aquatic life.
 Human Health
   For the maximum protection of human
 health from the potential carcinogenic
 effects due to exposure of toxaphene
 through tngesHon of contaminated water
 dnd contaminated aquatic organisms,
 the ambient water concentration should
 be zero based on the non-threshold
 assumption for this chemical However,
 zero level may not be attainable at the
 present time. Therefore, the levels which
 may result In Incremental increase of
 cancer risk over the lifetime are
 estimated at lO"8, W*. and 10"*. The
 corresponding criteria are 7.1 ng/1. .71
 r.g/L and JO/7 ng/1, respectively. If the
 above estimates are made for
 consumption of aquatic organisms only,
 excluding consumption of water, the
 levels are 7.3 ng/1, .73 ng/1, and IS! ng/1,
 respectively. Other concentrations
 representing different risk levels may be
 calculated by use of the Guidelines. The
 risk estimate range is presented for
 information purposes and does not
 represent an Agency judgment on an
 "acceptable" risk level.
 Trichloroethylene
 Freshwater Aquatic Life
  The available data for
 trlchloroethylene Indicate that acute
 toxicity to freshwater aquatic life occurs
 at concentrations as low as 45,000 jig/1
 and would occur at lower
 concentrations among species that are
 more sensitive than those tested. No
 data are available concerning the
 chronic toxicity of trichloroethylene to
 sensitive freshwater aquatic life but
 adverse behavioral effects occurs to one
 species at concentrations as low as
 21,900 tig/I.
 Saltwater Aquatic Life
  The available data for
 trichloroethylene indicate that acute
toxicity to salrwater'aquatic life occurs'
at concentrations as low as 2,000 pg/1
and would occur at lower
concentrations among species that are
more sensitive than those tested. No
data are available concerning the
chronic toxicity of trichloroethylene to
 sensitive saltwater aquatic life.
Human Health
  For the maximum protection of human
health from the potential carcinogenic
effects due to exposure of
trichloroethylene through ingeation of
 contaminated water and contaminated
aquatic organisms, the ambient water
concentration should be zero based on
 the non-threshold assumption for this
 chemical. However, zero level may not
 be attainable at the present time.
 Therefore, the levels which may result in
 incremental increase of cancer risk over
 the lifetime are estimated at 10"*, 10"«,
 and 10~T. The corresponding criteria are
 27 jig/L 2.7 Mg/1. and .27 jig/1,
 respectively. If the above estimates are
 made for consumption of aquatic
 organisms only, excluding consumption
 of water, the levels are 807 jtg/1, 80.7
 fig/1, and 8.07 jig/l respectively. Other
 concentrations representing different
 risk levels may be calculated by use of
 the Guidelines. The risk estimate range
 is presented for information purposes
 and does not represent an Agency
 Judgment on an "acceptable" risk level.

 Vinyl Chloride
 Freshwater Aquatic Life
   No freshwater organisms have been
 tested with vinyl chloride and no
 statement can be made concerning acute
 or chronic toxicity.
 Saltwater Aquatic Life
   No saltwater organisms have been
 tested with vinyl chloride and no
 statement can be made concerning acute
 or chronic toxicity.
 Human Health
   For the maximum protection of human
"health from the potential carcinogenic
 effects due to exposure of vinyl chloride
 through ingestion of contaminated water
 and contaminated aquatic organisms,
 the ambient water concentration should
 be zero based on the non-threshold
 assumption for this chemical However,
 zero level may not be attainable at the
 present time. Therefore, the levels which
 may result in incremental increase of
 cancer risk over the lifetime are
 estimated at 10~*, 10"B. and «rT. The
 corresponding criteria are 20 jig/1, 2.0
 jig/1, and ,2 jig/1, respectively. If the
 above estimates are made for
 consumption of aquatic organisms only,
 excluding consumption of water, the
 levels are 5,248 ug/1,525 MS/1, and 52.5
 fig/1, respectively. Other concentrations
 representing different risk levels may be
 calculated by use of the Guidelines. The
 risk estimate range is presented for
 information purposes and does not
 represent an Agency judgment on an
 "acceptable" risk level.
 Zinc

Freshwater Aquatic Life
   For total recoverable zinc the criterion
 to protect freshwater aquatic life as
 derived using the Guidelines is 47 jig/1
 as a 24-hour average and the
 concentration (in u.g/1] should not
exceed the numerical value given by
e
-------
 79342
Federal Register / Vol. 45. No. 231  / Friday. November  28. 1980 / Notices
 necessarily all of the species all of the
 time. Aquatic communities can tolerate
 some stress and occasional adverse
 effects on a few species, and so total
 protection of all of the species all of the
 time is not necessary. Rather, the
 Guidelines attempt to provide a
 reasonable and adequate amount of
 protection with only a small possibility
 of considerable overprotection or
 underprotection. Within these
 constraints, it seems appropriate to err
 on the side of overprotection.
  The numerical aquatic life criteria
 derived using the Guidelines are
 expressed as two numbers, rather than
 the traditional one number, so that the
 criteria can more accurately reflect
 lexicological and practical realities. The
 combination of both a maximum value
 and a 24-hour average value is designed
 to provide adequate protection of
 aquatic life and its uses from acute and
 chronic toxiclty to animals, toxicity to
 plants and bioconcentration by aquatic
 organisms without being as restrictive
 as a one-number criterion would have to
 be to provide the same amout of
 protection. The only way to assure the
 same degree of protection with a one-
 number criterion would be to use the 24-
 hour average as a concentration that is
 not to be exceeded at  any time in any
 place.
  The two-number criterion is intended
 to identify an average pollutant
 concentration which will produce a
 water qualtly generally suited to the
 maintenance of aquatic life and its uses
 while restricting the extent and duration
 of excursions over the average so that
 the total exposure will not cause
unacceptable adverse effects. Merely
specifying an average value over a time
period is insufficient, unless the period
of time is rather short, because of
concentration higher than the average
value can kill or cause substantial
damage in short periods. Furthermore,
for some substances the effect of
intermittent high exposures is
cumulative. It is therefore necessary to
place an upper limit on pollutant
concentrations to which aquatic
organisms might be exposed, especially
when the maximum value is not much
higher than the average value. For some
substances the maximum may be so
much higher than the 24-hour average
that in any real-world  situation the
maximum will never be reached if the
24-hour average is achieved. In such
cases the 24-hour average will be
limiting and the maximum will have no
practical significance,  except to indicate
that elevated concentrations are
acceptable as long as the 24-hour
average is achieved.
                       These Guidelines have been
                     developed on the assumption that the
                     results of laboratory tests are generally
                     useful for predicting what will happen in
                     field situations. The resulting criteria are
                     meant to apply to most bodies of water
                     in the United States, except for the
                     Great Salt Lake. All aquatic organisms
                     and their common uses are meant to be
                     considered, but not necessarily
                     protected, if relevant data are available,
                     with at least one specific exception. This
                     exception is the accumulation of
                     residues of organic compounds in the
                     siscowet subspecies of lake trout which
                     occurs in Lake Superior and contains up
                     to 67!5 fat in the fillets (Thurston, C.E..
                     1962, Physical Characteristics and
                     Chemical Composition of Two
                     Subspecies of Lake Trout, J. Fish. Res.
                     Bd. Canada 19:39-44). Neither siscowet
                     nor organisms in the Great Salt Lake are
                     intentionally protected by these
                     Guidelines because both may be too
                     atypical.
                      . With appropriate modifications these
                     Guidelines can be used to derive criteria
                     for any specified geographical area,
                     body of water (such as the Great Salt
                     Lake), or group of similar bodies of
                     water. Thus With appropriate
                     modifications the Guidelines can be
                     used to derive national, state, or local  '
                     criteria if adequate information is
                     available concerning the effects of the
                     substance of concern on appropriate
                     species and their uses. However, the
                     basic concepts described in the
                     Guidelines should be modified only
                     when sound scientific evidence
                     indicates that a criterion produced using
                     the Guidelines would probably
                     significantly overprotect or underprotect
                     the presence or uses of aquatic life.
                       Criteria produced by these Guidelines
                     are not enforceable numbers. They may
                     be used in developing enforceable
                     numbers, such as water quality
                     standards and effluent standards.
                     However, the development of standards
                     may take into account additional factors
                     such as social, legal, economic, and
                     hydrologies! considerations, the
                     environmental and analytical chemistry
                     of the substance, the extrapolation from
                     laboratory data to field situations, and
                     the relationship between the species for
                     which data are available and the
                     species which are to be protected.
                       Because fresh water and salt water
                     (including both estuarine and marine
                     waters) have basically different
                     chemical compositions and because
                     freshwater and saltwater species rarely
                     inhabit the same water simultaneously,
                     separate criteria should be derived for
                     these two kinds of waters. However, for
                     some substances sufficient data may not
be available to allow derivation of one
or both of these criteria using the
Guidelines.
  These Guidelines are mean* to be
used after a decision is made that a
criterion is needed for a substance. The
Guidelines do not address the rationale
for making that decision. If the potential
for adverse effects on aquatic life and
its uses are part of the basis for deciding
whether or not a criterion is needed for
a substance, these Guidelines may be
helpful in the collection and
interpretation of relevant data.
/. Define the  Substance for Which the
Criterion Is To Be Derived
  A. Each separate chemical which
would not ionize significantly in most
natural bodies of water should usually
be considered a separate substance,
except possibly for structurally similar
organic compounds that only differ in
the number and location of atoms of a
specific halogen, and only exist in large
quantities as commercial mixtures of the
various compounds, and apparently
have similar chemical, biological, and
toxicological properties.
  B. For chemicals, which would ionize
significantly in most natural bodies of
water, such as inorganic salts, organic
acids and phenols, all forms that would
be in chemical equilibrium should
usually be considered one substance.
For metals, each different valence and
each different covalently bonded
organometallic compound should
usually be considered a separate
substance.
  C. The definition of the substance may
also need to take into account the
analytical chemistry and fate of the
substance.
//. Collect and Review A vailable Data
  A. Collect all available data on the
substance concerning (1) toxicity to, and
bioaccumulation by. aquatic animals
and plants, (2) FDA action levels, and
(3) chronic feeding studies with wildlife.
  B. Discard all data that are not
available in hard copy (publication,
manuscript, letter, memorandum, etc.)
with enough supporting information to
indicate that acceptable test procedures
were used and that the results are
reliable. Do not assume that all
published data are acceptable.
  C. Discard questionable data. For
example, discard data from tests for
which no control treatment existed, In
which too many organisms in the control
treatment died or showed signs of stress
or disease, or in which distilled or
deionized water was used as the
dilution water for aquatic organisms.
Discard data on formulated mixtures
and emulsifiable concentrates of the

-------
                   Federal Register  /  .Vol. 45. No. 231  /  Friday.  November 28.  1980 /  Notices             79343
 substance of concern, but not
 necessarily data on technical grade
 material.
  D. Do not use data obtained using;
  1. Brine shrimp, because they usually
 only occur naturally in water with
 salinity greater than 35 g/kg.
  2. Species that do not nave
 reproducing wild populations resident
 in—but not necessarily native to—North
 America. Resident North American
 species of fishes are defined as those
 listed In "A List of Common and
 Scientific Names of Fishes from the
 United States and Canada", 3rd ed.,
 Special Publication No. 6, American
 Fisheries Society, Washington, D.C.,
 1970. Data obtained with non-resident
 species can be used to Indicate
 relationships and possible problem
 areas, but cannot be used in the
 derivation of criteria.
  9, Organisms that were previously
 exposed to significant concentrations of
 the test material or other pollutants.
 ///.  Minimum Data Base
  A, A minimum amount of data should
 be available to help ensure that each of
 the four major kinds of possible adverse
 effects receives some consideration.
 Results of acute and chronic toxicity
 tests with a reasonable number and
 variety of aquatic animals are necessary
 so that data available for tested species
 can be considered a useful indication of
 the sensitivities of the numerous
 untested species. The requlurements
 concerning toxicity to aquatic plants  are
 less stringent because procedures for
 conducting tests with plants are not as
 well developed and  the interpretation of
 the  results is more questionable. Data
concerning bioconcentration by aquatic
organisms can only be used if other
relevant data are available.
  B. To derive a criterion for freshwater
aquatic life, the following should be
available:
  1. Acute tests (see Section IV) with
freshwater animals in at least eight
 different families provided that of the
eight species;
—at least one is a salmonid fish
—at least one is a non-sabnonid fish
—at least one is a planktonic crustacean
—at least one Is a benthic crustacean
—at least one is a benlhlc insect
—at least one of the benthic species is a
  detritivore
  2, Acute-chronic ratios (see Section
VI]  for at least three species of aquatic
animals provided that of the three
species:
—at least one is a fish
—at least one is an invertebrate
—at least one is a freshwater species
  (the other two may be  saltwater
  species]
  3. At least one test with a freshwater
alga or a chronic test with a freshwater
vascular plant (see Section VIII). If
plants are among the aquatic organisms
that are most sensitive to the substance,
tests with more than one species should
be available.
  4. At least one acceptable
bioconcentration factor determined with
an aquatic animal species, if a maximum
permissible tissue concentration Is
available (see Section DC).
  C. To derive a criterion for saltwater
aquatic life, the following should be
available;
  1. Acute tests (see Section IV) with
saltwater animals in at least eight
different families provided that of the
eight-species:
—at least two different fish families are
  included
—at least five different invertebrate
  families are included
—either the Mysldae or Penaeidae
  family or both are included
—at least one of the Invertebrate
  families is in a phylum other than
  Arthropoda
  2. Acute-chronic ratios (see Section
VI} for at least three species of aquatic
animals provided that of the three
species:
—at least one is a fish
—at least one is an invertebrate
—at least one is a saltwater species (the
  other two may be freshwater species]
  3. At least one test with a saltwater
alga or a chronic test with a saltwater
vascular plant (see Section VIII]. If
plants are among the aquatic organisms
most sensitive to the substance, tests
with more than one species should be
available.
  4. At least one acceptable
bioconcentration factor determined with
an aquatic animal species, if a maximum
permissible tissue concentration Is
available (see Section DC).
  D. If all the requirements of the
minimum data base are met, a criterion
can usually be derived, except in special
cases. For example, a criterion might not
be possible if the. acute-chronic ratios
vary greatly with no apparent pattern.
Also,  if a criterion is to be related to a
water quality characteristic, (see
Sections V and VII), more data will be
necessary.
  Similarly, if the minimum data
requirements are not satisfied, generally
a criterion should not be derived except
in special cases. One such special case
would be when less than the minimum
amount of acute and chronic data are
available, but the available data clearly
indicate that the Final Residue Value
would be substantially lower then  either
the Final Chronic Value or the Final
Plant  Value.
 IV. Final Acute Value
   A. Appropriate measures of the acute
 (short-term) toxicity of the substance to
 various species of aquatic animals are
 used to calculate the Final Acute Value.
 If acute values are available for fewer
 than twenty species, the Final Acute
 Value probably should be lower than
 the lowest value. On the other hand, if
 acute values are available for more than
 twenty species, the Final Acute Value
 probably should be higher than the
 lowest value, unless the most sensitive
 species is an important one. Although
 the procedure used to calculate the Final
 Acute Value has some limitations. It
 apparently Is the best of the procedures
 currently available.
   B. Acute toxicity tests should be
 conducted using procedures such as
 those described In:
   ASTM Standard E 729-80, Practice for
 Conducting Acute Toxicity Tests with
 Fishes, Macroinvertebrates, and
 Amphibians. American Society for
 Testing and Materials, 1916 Race Street,
 Philadelphia, PA 18103.
   ASTM Standard 1724-40, Practice for
 Conducting Static Acute Toxicity Tests
 with Larvae of Four  Species of Bivalve
 Molluscs. American Society for Testing
 and Materials. 1916 Race Street,
 Philadelphia. PA 19103.
   C. Results  of acute tests in which food
 was added to the test solutions should
 not be used, because this may
 unnecessarily affect the results of the
 test.
   D. Results  of acute tests conducted
 with embryos should not be used (but
 see Section FV.E.2), because this is often
 an Insensitive life stage.
   E. Acute values should be based on
 endpoints and lengths  of exposure
 appropriate to the life stage of the
 species tested. Therefore, only the
 following kinds of data on acute toxicity
. to aquatic animals should be used:
   1.48-hr EC50 values based on
 Immobilization and 48-hr LC50 values
 for flrst-instar (less than 24 hours old)
 daphnlds and other cladocerans, and
 second- or thlrd-instar midge larvae.
   2.48- to 96-hr ECSO values based on
 Incomplete shell development and 48- to
 96-hr LC50 values for embryos and
 larvae of barnacles,  bivalve molluscs
 (clams, mussels, oysters, and scallops],
 sea urchins, lobsters, crabs, shrimps,
 and abalones.
   3.96-hr ECSO values based on
 decreased shell deposition for oysters.
   4.96-hr ECSO values on
 immobilization or loss of equilibrium or
 both and 96-hr LCSO values for aquatic
 animals, except for cladocerans, midges,
 and animals whose behavior or
 physiology allows .them to avoid

-------
 79344
Federal Register / Vol. 45. No. 231  / Friday, November 28. 1980 / Notices
 exposure to toxicant or for whom the
 acute adverse effect of the exposure
 cannot be adequately measured. Such
 freshwater and saltwater animals
 include air-breathing molluscs, unionid
 clams, operculate snails, and bivalve
 molluscs, except for some species that
 cannot "close up" and thus prevent
 exposure to toxicant, such as the bay
 scallop (Aigppecten irradians).
  V. For the 'use of LC50 or EC5Q values
 for durations shorter and longer than
 those listed above, see Section X
  G. If the acute toxic! ty of the
 substance to aquatic animals has been
 shown to be related to a water quality
 characteristic such as hardness for
 freshwater organisms or salinity for
 saltwater organisms, a Final Acute
 Equation should be derived based on
 that water quality characteristic. Go to
 Section V.
  H. If the acute toxicity of the
 substance has not been adequately
 shown to be related to a water quality
 characteristic, for each species tor
 which at least one acute value is
 available, calculate the geometric mean
 of the results of all flow-through tests in
 which the toxicant concentrations were
 measured. For a species for which no
 such result is available, calculate the
 geometric mean of all available acute
 values, i.e., results of flow-through tests
 in which the toxicant concentrations
 were not measured and results of static
 and renewal testa based on initial total
 toxicant concentrations.
  Not*.—The geometric mean of N numbers
 te obtained by taking the N«* root of the
 product of N numbers. Alternatively, the
 geometric mean can be calculated  by adding
 the logarithms of the N numbers, dividing the
 sum by N, and taking the antilog of the
 quotient The geometric mean, of two numbers
 con alia be calculated as the square loot of
 the product of the two numbers. The
 geometric mean of one number is that
 number. Either natural (base a) or common
 (base 10} logarithms can be used to calculate
 geometric means as long as they are used
 consistently within each.set of data, Le« the
 antilog used must match the logarithm used.
  I. Count the number «N of species for
which a species  mean acute value is
 available.
  J. Order the species mean acute
 values from low to high. Take the
 common logarithms of the N values (log
 mean values).
  K. The intervals (cell widths) for the
 lower cumulative proportion
 calculations are  0.11 common log units
 apart, starting from the lowest  log value.
The value of 0.11 is an estimate of
 average precision and was calculated
 from replicate species acute values.
  L. Starting with the lowest log mean
 value, separate the N values into
                     intervals (or cells) calculated in Step IV,
                     K.
                       M. Calculate cumulative proportions
                     for each non-empty interval by summing
                     the number of values In the present and
                     all lower intervals and dividing by N.
                     These calculations only need to be done
                     for the first three non-empty intervals
                     (or cells).
                       N. Calculate the arithmetic mean of
                     the log mean values for each of the three
                     intervals.
                       O. Using the two interval mean acute
                     values and cumulative proportions
                     closest to 0.05, linearly extrapolate or
                     Interpolate to the O.OS log concentration.
                     The Final Acute Value is the antilog of
                     the O.OS concentration. ~
                       In other words, where
                     Prop(l) and conc(l] are the cumulative
                       proportion, and mean log value for the
                       lowest non-empty interval.
                     Prop(2) and conc(2) are the cumulative
                       proportion and mean log value for the
                       second lowest non-empty interval
                     A <= Slope of the cumulative proportions
                     B=The 0.05 log value
                     Them
                     A=[0.05-Prop(l]]/[Prop{2)-Prop(l]]
                     B=conc(l)+A |conc(2)-conc(l)J
                     Final Acute Value=10s
                       P. If for an important species, such as
                     a recreationally or commercially
                     important species, the geometric mean
                     of the acute values from flow-through
                     tests m which the toxicant
                     concentrations were measured is lower
                     than the Final Acute Value, then that
                     geometric mean should be used as the
                     Final Acute Value.
                       Q. Go to Section VL

                     V, Final Acute Equation
                      A. When enough data are available to
                     show that acute toxicity to two ot more
                     specie&ls similarly affected by a water
                     quality characteristic, this effect can be
                     taken into account as described below.
                     Pooled regression analysis should
                     produce similar results, although data
                     available for individual species would
                     be weighted differently.
                      B. For each species for which
                     comparable acute toxicity values are
                     available at two or more different
                     values of a water quality characteristic
                     which apparently affects toxicity,
                     perform a least squares regression of the
                     natural logarithms of the acute toxicity
                     values on the natural logarithms of the
                     values of the water quality
                     characteristic. (Natural logarithms
                     [logarithms to the base e, denoted as In]
                     are used herein merely because they are
                     easier to use on some hand calculators
                     and computers than common logarithms
                     [logarithms to the base 10].  Consistent
                     use of either will produce the same
result.) No transformation or a different
transformation may be used'if it fits the
data better, but appropriate changes will
be necessary throughout this section.
   C. Determine whether or not each
acute slope Is meaningful, taking into
account the range and number of values
of the water quality characteristic
tested. For example, a slope based on
four data points may be of limited value
if it is based only on data for a narrow
range of values of the water quality
characteristic. On the other hand, a
slope based on only two data points
may be meaningful if it is consistent
with other information and if the two
points cover a broad enough range of  •
the water quality characteristic. If
meaningful slopes are not available for
at least two species or if the available
slopes are not similar, return to Section
IV. H.. using the results of tests
conducted under conditions-and in
•water similar to those commonly used
for toxicity tests with the species.
   D. Calculate the mean acute slope [V)
as the arithmetic average of all the
meaningful acute slopes for individual
species.
   E. For each species calculate the
geometric mean (W) of the acute toxicity
values and the geometric mean (XJ of
the related values of the water quality
characteristic,
   F. For each species calculate the
logarithmic intercept (Y) using the
equation: Y=ta W-V(ln X).
   G. For each species calculate the
species mean acute intercept as the
antilog of Y.
   H. Obtain the Final Acute Intercept by
using the procedure described in Section
IV. I-O, except Insert "Intercept" for
"Value".
   I. If for an important species, such as a
recreationally or commercially
important species, the intercept
calculated only from results of flow-
through tests in which the toxicant
concentrations were measured is lower
than the Final Acute Intercept, then that
intercept should be used as the Final
Acute Intercept.
   J. The Final Acute Equation Is written
as e'*1""'"1" i""" eh««euii«u«a+i» »t where
V=mean acute slope and Z=Final
Acute Intercept
VI Final Chronic Value
   A. The Final Chronic Value can be
calculated in the same manner as the
Final Acute Value or by dividing the
Final Acute Value by the Final Acute*
Chronic Ratio, depending on the data
available. In some cases it will not be
possible to calculate a Final Chronic
Value.
   B. Use only the results of flow-through
(except renewal is acceptable for

-------
                   Federal Register / Vol. 45. No.  231 / Friday. November  28. 1880 / Notices     	79315
daphnids) chronic tests in which the
concentrations of toxicant in the test
solutions were measured.
  C. Do not use the results of any
chronic test In which survival, growth,
or reproduction among the controls was
unacceptably low.
  D. Chronic values should be based on
endpoints and lengths of exposure
appropriate to the species. Therefore,
only the results of the following kinds of
chronic toxlclty tests should be used:
  1. Life-cycle toxicity tests consisting
of exposures of each of several groups
of individuals of a species to a different
concentration of the toxicant throughout
a life cycle. To ensure that-all life stages
and life processes are exposed, the test
should begin with embryos or newly
hatched young less than 48 hours old
(iess than 24 hours old for daphnids],
continue through maturation and
reproduction, and with fish should end
not less than 24 days (80 days for
salmonids) after the hatching of the next
generation. For fish, data should be
obtained and analyzed on survival and
growth of adults arid young, maturation
of males and females, embryos spawned
per female, embryo viability (salmonids
only) and hatchability. For daphnids,
data should be obtained and analyzed
on survival and  young per female.
  2. Partial life-cycle toxidty tests
consisting of exposures of each of
several groups of individuals of a
species of fish to a different
concentration of the toxicant through
uost portions of a life cycle. Partial life-
:ycle tests are conducted with fish
species that require more than a year to
each sexual maturity, so that the test
:an be completed In less than 15
aonths, but still expose all major life
•tages to the toxicant. Exposure to the
oxicant begins with immature Juveniles
it least 2 months prior to active gonad
levelopment, continues through
naturation and reproduction, and ends
tot  less than 24  days (90 days for
almonids) after the hatching of the next
eneration. Data should be obtained and
 nalyzed on survival and growth of
 dults and young, maturation of males
nd females, embryos spawned per
•male, embryo viability (salmonids
nly] and hatchability.
  3. Early-life-stage toxicity tests
 insisting of 28- to 32-days (60 days
 Dst-hatch for salmonids] exposures of
 ie early life stages of a species of fish
 om shortly after fertilization through
 nbryonic, larval, and early juvenile
 jvelopment. Data should be obtained
 id analyzed on survival and growth.
  E. Do not use the results of an tarly-
  s-stage test if results of a life-cycle or
 irtial life-cycle test with the same
,  -ecies are available.
  F, A chronic value is obtained by
 calculating the geometric mean of the
 lower and upper chronic limits from a
 chronic test. A lower chronic limit is the
 highest tested concentration (1) in an
 acceptable chronic test, [2] which did
 not cause the occurrence (which was
 statistically significantly different from
 the control at p-0.05] of a specified
 adverse effect, and (3) below which no
 tested concentration caused  such an
 occurrence. An upper chronic limit is the
 lowest tested concentration (I) in an
 acceptable chronic test, (2] which did
 cause the occurrence (which was
 statistically significantly different from
 the control at p=0.05) of a specified
 adverse effect and (3) above which all
 tested concentrations caused such an
 occurrence.
  Note.—Various authors have used a
 variety of terms and definitions to Interpret
 the results of chronic tests, so reported
 results should be reviewed carefully.
  G. If the chronic toxicity of the
 substance to aquatic animals has been
 adequately shown to be related to a
 water quality characteristic such as
 hardness for freshwater organisms or
 salinity for saltwater organisms, a Final
 Chronic Equation should be derived
based on that water quality
characteristic. Go to Section VII.
  H. If chronic values are available for
eight species as described In Section 01.
B.I or m. C.I, a species mean .chronic
value should be calculated for each
 species for which at least one chronic
value is available by calculating the
geometric mean of all the chronic values
for the species. The Final Chronic Value
 should then be obtained using the
procedures described in Section IV. I-O.
Then go to Section VI. M.
  I. For each chronic value for which at
least one appropriate acute value is
available, calculate an acute-chronic
ratio, using for the numerator the
arithmetic average of the results of alJ
standard flow-through acute tests in
which the concentrations were
measured and which are from the same
study as the chronic test. If such an
acute test is not available, use for the
numerator the results of a standard
acute test performed at the same
laboratory with the same species,
toxicant and dilution water. If no such
acute test is available, use the species
mean acute value for the numerator.
  Note.—If the acute toxlclty or chronic
toxicity or both of the substance have been
adequately shown to be related to a water
quality characteristic, the numerator and the
denominator must be based on tests
performed in the same water.
  J. For each species, calcuate the
species  mean acute-chronic ratio as the
 geometric mean of all the acute-chronic
 ratios available for that species.
   K. For some substances the species
 mean acute-chronic ratio seems to be
 the same for all species, but for other
 substances the ratio seems to increase
 as the species mean acute value
 increases. Thus the Final Acute-Chronic
 Ratio can be obtained in two ways,
 depending on the data available.
   1, If no major trend is apparent and
 the acute-chronic ratios for a number of
 species are within a factor of ten, the
 final Acute-Chronic Ratio should be
 calculated as, the geometric mean of all
 the species mean acute-chronic ratios
 available for both freshwater and
 saltwater species.
   2. If the species mean acute-chronic
 ratio seems to increase as the species
 mean acute value increases, the value of
 the acute-chronic ratio for species
 whose acute values are close to the
 Final Acute Value should be chosen as
 the Final  Acute-Chronic Ratio.
   L Calculate the  Final Chronic Value
 by dividing the Final Acute Value by the
 Final Acute-Chronic Ratio.
   M. If the species mean chronic value
 of an important species, such as a
 commercially or recreationally
 important species, is lower than the
 Final Chronic Value, then that species
 mean chronic value should be used as
 the Final  Chronic Value.
   N. Go to Section VIII.
 VII. Final Chronic Equation
  . A. For each species for which
 comparable chronic toxicity values are
 available at two or more different
 values of a water quality characteristic
 which apparently affects chronic
 toxicity, perform a least squares
 regression of the natural-logarithms of
 the chronic toxidty values on the
 natural logarithms of the water quality
 characteristic values. No transformation
 or a different transformation may be
 used if it fits the data better, but
 appropriate changes will be necessary
'throughout this section. It is probably
 preferable, but not necessary,-to use the
 same transformation that was used with
 the acute values in Section V,
   B. Determine whether or not each
 chronic slope is meaningful, taking into
 account the range and number of values
 of the water quality characteristic
 tested. For example, a slope based on
 four data points may be of limited value
 if it is based only on data for a narrow
 range of values of the water quality
 characteristic. On  the other hand, a .
 slope based on only two  data points
 may be meaningful if it is consistent
 with other information and if'the two
 points cover a broad enough range of
 the water quality characteristic. If a

-------
79348
Federal  Register / Vol. 45, No. 231  / Friday, November  28.  1980 / Notices
meaningful chronic slope is not
available for at least one species, return
to Section VI. H.
  C. Calculate the mean chronic slope
(L) as the arithmetic average of all the
meaningful chronic slopes for individual
species.
  D. For each species calculate the
geometric mean (M) of the toxicity
values and the geometric mean (P) of the
related values of the water quality
characteristic.
  E. For each species calculate the
logarithmic intercept (Q) using the
equation: Q=lnM-L(ln P).*
  F. For each species calculate a species
mean chronic intercept as the antilog of
Q.  -
  G. Obtain the Final Chronic Intercept
by using the procedure described in
Section TV. I-O, except insert
"Intercept" for "Value".
  H. If the species mean chronic
intercept of an important species, such
as a commercially or recreationally
important species. Is lower than the
Final Chronic Intercept then that
species mean chronic intercept should
be used as the Final,Chronic Intercept.
  I. The Final Chronic Equation is
written as e Q&*t**t*T.n*utr d»>«Mri«uc>|+tn
B), where L=mean chronic slope and
R= Final Chronic Intercept
VIII. Final Plant Value
  A. Appropriate measures of the
toxicity of the substance to aquatic
plants are used to compare  the relative
sensitivities of aquatic plants and
animals.
  B. A value is a concentration which
decreased growth (as measured by dry
weight chlorophyll etc.) in  a 96-hr or
longer test with an alga or in a chronic
test with, an aquatic vascular plant
  C. Obtain the Final Plant Value by
selecting the lowest plant value from a
test in which the toxicant concentrations
were measured.
IX, Final Residue Value
  A. The Final Residue Value is derived
in order to (1] prevent commercially or
recreationally important aquatic
organisms from exceeding relevant FDA
action levels and (2) protect wildlife,
including fishes and birds, that eat
aquatic organisms from demonstrated
adverse effects. A residue value ia
calculated by dividing a maximum
permissible tissue concentration by an
appropriate bioconcentration factor
(BCF). where the BCF is the quotient of
the concentration of a substance in all
or part of an aquatic organism divided
by the concentration in water to which
the organism has been exposed. A
maximum permissible tissue
concentration is either (1) an action
                     level from the FDA Administrative
                     Guidelines Manual for fish oil or for the
                     edible portion of fish or shellfish, or [2] a
                     maximum acceptable dietary intake
                     based on observations on survival,
                     growth or reproduction in a chronic
                     wildlife feeding study. If no maximum
                     permissible tissue concentration is
                     available, go to Section X because no
                     Final Residue Value can be derived.
                       B. L A BCF determined in a
                     laboratory test should be used only if it
                     was calculated based on measured
                     concentrations of the substance in the
                     test solution and was based on an
                     exposure that continued until either
                     steady-state or 28-days was reached.
                     Steady-state is-reached when the BCF
                     does not change significantly over a
                     period of time, such as two days or 16
                     percent of the length of the exposure,
                     whichever is longer. If a steady-state
                     BCF is not available for a species, the
                     available BCF for the longest exposure
                     over 28 days should be used for that
                     species.
                       2. A BCF from a field exposure should
                     be used only when it is known that the
                     concentration, of the substance was
                     reasonably constant for a long enough
                     period of time over the range of territory
                     inhabited by the organisms.
                       3. If BCF values from field exposures
                     are consistently lower or higher than
                     those from laboratory exposures, then
                     only those values from field.exposures
                     should be used if possible.
                       4. A BCF should be calculated based
                     on the concentration of the substance
                     and its. metabolites, which are
                     structurally similar and are not much1
                     more soluble in water than the parent
                     compound, in appropriate tissue and
                     should be corrected for the
                     concentration in the organisms at the
                     beginning of the test.
                       5. A BCF value obtained from a
                     laboratory or field exposure that caused
                     an observable advene effect on the test
                     organism may be used only if it is
                     similar to that obtained with unaffected
                     organisms at lower concentrations in the
                     sametesL
                       6. Whenever a BCF is determined for
                     a lipid-soluble substance, the percent
                     lipids should also be determined in the
                     tissue for which the BCF was calculated.
                       C. A BCF calculated using dry tissue
                     weights must be converted to a wet
                     tissue weight basis by multiplying the
                     dry weight BCF value by 0.1 for
                     plankton and by 0.2 for individual
                     species of fishes and invertebrates.
                       Note.—The values of 0.2 and 0.1 were
                     derived from data published in:
                     McDiffett, W. F., 1870, Ecology 51:875-988.
                     Bracken. R. W, et aL 1988. J. Wildlife
                       Management 32:52-75.
Cummins. K. W., et al. 1973. Ecology 54: 338-
  345.
Pesticide Analytical Manual. Volume 1, Food
  and Drug Administration. 1988. •
Love. R. M.. 1957. In The Physiology of Fishes,
  Vol. I. M. E. Brown, ed. Academic Press
  New York. p. 411.
Ruttner, F.. 1963. Fundamentals of Limnology.
  3rd ed. Trans, by D. G. Frey and F. E. |. Fry.
  Univ. of Toronto Press, Toronto.
  Some additional values can be found in:
Sculthorpe. C. D., 1987. The Biology of
  Aquatic Vascular Plants. Arnold Publishing
  Ltd., London.
  D. If enough 'pertinent data exist
several residue values can be calculated
by dividing maximum permissible tissue
concentrations by appropriate BCF
values.
  1. For each available maximum.
acceptable dietary intake derived from a
chronic feeding study wish wildlife,
including birds and aquatic  organisms.
the appropriate BCF is based on the
whole body of aquatic species which
constitute or represent a major portion
of the diet of the tested wildlife species.
  2. For an FDA action level, the
appropriate BCF is the highest geometric
mean species BCF for the edible, portion
(muscle for decapods- muscle with or
without skin .for fishes, adductor muscle .
for scallops and total living tissue for
other bivalve molluscs) of a consumed
species. The highest species BCF is used
because FDA action levels are applied
on a specles-by-species basis.
  E. For lipJd-soluble substances, it may
be possible to calculate additional
residue values. Because steady-state
BCF values for a lipid-soluble chemical
seem to be proportional to percent lipids
from one tissue to another and from one
species, to another, extrapolations can
be made from tested tissues or species
to untested tissues or species on the
basis of percent lipids.
  1. For each BCF for which the percent
lipids is known for the same tissue for
which the BCF was measured,  the BCF
should be normalized to a one percent
lipid basis by dividing the BCF by the
percent lipids. This adjustment to a one
percent lipid basis makes all the
measured BCF values comparable
regardless of the species or tissue for
which the BCF was measured.
  2. Calculate the geometric mean
normalized BCF. Data for both saltwater
and freshwater species can be used to
determine the mean normalized BCF.
because the normalized BCF seems to
be about the same for both  kinds of
organisms.
   3. Residue values can then be
calculated by dividing the maximum
permissible tissue concentrations by the
mean normalized BCF and by a percent
lipids value appropriate to the maximum
permissible tissue concentration, i.e..

-------
                  Federal Register / Vol. 45.  No. 231  / Friday.  November 28.  1930 / Notices
                                                                                                             79347
Residua Value • jnaximua  permissible  tissue  concentration)
                  (mean normalized BC7)(appropriate  percent  lipids)

                                       exceeded on the average in a 24-hour
                                       period and one that should not be
                                       exceeded at any time during the 24-hour
                                       period. This two-number criterion is
                                       intended to identify water quality
                                       conditions that should protect aquatic
                                       life and its uses from acute and chronic
                                       adverse effects of both cumulative and
                                       noncumulative substances without being
                                       as restrictive as a one-number criterion
                                       would have to be to provide the same
                                       degree of protection.
                                         B. The maximum concentration is the
                                       Final Acute Value or is obtained from
                                       the Final Acute Equation.
                                         C. The 24-hour  average concentration
                                       is obtained from the Final Chronic
                                       Value, the Final Plant Value, and the
                                       Final Residue Value by selecting the
                                       lowest available value, unless other
                                       data (see Section X) from tests in which
                                       the toxicant concentrations were
                                       measured show that a lower value
                                       should be used. If toxicity is related to a
                                       water quality characteristic, the 24-hour
                                       average concentration is obtained from
                                       the Final Chronic Equation, the Final
                                       Plant Value, and the Final Residue
                                       Value by selecting the one that results in
                                       the lowest concentrations in the normal
                                       range of the water quality characteristic,
                                       unless other data (see Section X] from
                                       tests in which the toxicant
                                       concentrations were measured show
                                       that a lower value should be used.
                                         D. The criterion is (the 24-hour
                                       average concentration) as a 24-hour
                                       average and the concentration should
                                       not exceed (the maximum
                                       concentration] at any time.

                                       XII. Review

                                         A. On the basis of all available
                                       pertinent laboratory and field
                                       information, determine if the criterion is
                                       consistent with sound scientific
                                       evidence. If it is not, another criterion,
                                       either higher or lower, should be derived
                                       using appropriate modifications of the
                                       Guidelines.
                                         These Guidelines were written by
                                       Charles E. Stephen, Donald I. Mount,
                                       David ]. Hansen, John H. Gentile, Gary
                                       A. Chapman and William A. Brungs of
                                       the U.S.E.P.A. Environmental Research
                                       Laboratories in Corvallis, Oregon,
                                       Duluth, Minnesota, Gulf Breeze.  Florida,
                                       and Narragansett, Rhode Island.
                                       Numerous other people, many of whom
                                       do not work for U.S.E.P.A.. provided
                                       assistance and suggestions.
  a. For an FDA action level for fish oil,
the appropriate percent lipids value is
100.
  b. For an FDA action level for fish, the
appropriate percent lipids value is 15 for
fr<  .water criteria and 16 for saltwater
cr. eria because FDA action levels are
applied on a species-by-species basis to
commonly consumed species. The edible
portion of the freshwater lake trout
averages about IS percent lipids, and
the edible portion of the saltwater
Atlantic herring averages about 16
percent lipids (Sidwell, V. D., et al. 1974
Composition of the Edible Portion of
Raw (Fresh or Frozen] Crustaceans,
Finfish, and Mollusks. L Protein. Fat,
Moisture, Ash, Carbohydrate, Energy
Value, and Cholesterol. Marine Fisheries
Review 36:21-35).
  c. Foe a maximum acceptable dietary
intake derived from a chronic feeding
study with wildlife, the appropriate
percent lipids is the percent lipids of an
aquatic species or group of aquatic
species which constitute a major portion
of the diet of the wildlife species.
  F. The Final Residue Value is
obtained by selecting the lowest of the
available residue values. It should be
noted that in many cases the Final
Residue Value will not be low enough.
For example, a residue value calculated
from an FDA action level would result in
an average concentration in the edible
portion of a fatty species that is at the
action level On the average half of the
individuals of the species would have
concentrations above the FDA action
level. Also, the results of many chronic
feeding studies are concentrations that
cause adverse effects.
X. Other Data

  Pertinent information that could not
be used in earlier sections may be
available concerning adverse effects on
aquatic organisms and their uses. The
most important of these are data on
flavor impairment, reduction in survival,
growth, or reproduction, or any other
adverse effect that has been shown to
be biologically significant. Especially
important are data for species for which
10 other data are available. Data from
Behavioral, micorcosm, field, and
Dhysiological studies may also be
ivailable.

.'/. Criterion

  A. The criterion consists of two
oncentrations, one that shquld not be
Appendix C-Guidellnes and
Methodology Used In the Preparation of
Health Effect Assessment Chapters of
the Consent Decree Water Criteria
Documents
/ Objective
  The objective of the health effect
assessment chapters of the ambient
water criteria documents is to estimate
ambient water concentrations which do
not represent a significant risk to the
public. These assessments should
constitute a review of all relevant
information on individual chemicals or
chemical classes in order to derive
criteria that represent, in the case of
suspect or proven carcinogens, various
levels of incremental cancer risk, or, in
the case of other pollutants, estimates of
no-effect levels.
  Ideally, ambient water quality criteria
should represent levels for compounds
in ambient water that do not pose a
hazard to the human population.
However, in any realistic assessment of
human health hazard, a fundamental
distinction must be made between
absolute safety and the recognition of
some risk. Criteria for absolute safety
would have to be based on detailed
knowledge of dose-response
relationships in humans, including all
sources of chemical exposure, the types
of toxic effects elicited, the existence of
thresholds for the toxic effects, the
significance of toxicant interactions, and
the variances of sensitivities and
exposure levels within the human
population. In practice, such absolute
criteria cannot be established because
of deficiencies in both the available data
and the means of interpreting this
information. Consequently, the
Individual human health effects chapters
propose criteria which minimize or
specify the potential risk of adverse
human effects due to substances in
ambient water. Potential social or
economic costs and benefits are not
considered in the formulation of the
criteria.
II. Types of Criteria
  Ambient water quality criteria are
based on three types of biological
endpoints: carcinogenicity, toxicity (i.e.,
all adverse effects other than cancer),
and organoleptlc effects.
  For the purpose of deriving ambient
water quality criteria, carcinogenicity is
regarded as a non-threshold
phenomenon. Using this assumption,
"safe" or "no effect" levels for
carcinogens cannot be established'
because even extremely small doses
must be assumed to elicit a finite
increase in the incidence of the
response. Consequently, water quality

-------
79348
Federal  Register / Vol. 45. No. 231  /  Friday.  November 28.J19BO / Notices
criteria for carcinogens are presented as
a range of pollutant concentrations
associated with corresponding
incremental risks.
  For compounds which do not manifest
any apparent carcinogenic effect the
threshold assumption is used in deriving
a criterion. This assumption is based on
the premise that a physiological reserve
capacity exists within the organism
which is thought to be depleted before
clinical disease ensues. Alternatively, it
may be assumed that the rate of damage
will be Insignificant over the life span of
the organism. Thus, ambient water .
quality criteria are derived for non-
carcinogenic chemicals, and presumably
result in no observable-adverse-effect
levels (NOAELs) in the exposed human
population.
  In some instances, criteria are based
on organoleptiO'Characteristics, i.e.,
thresholds for taste or odor. Such
criteria are established when
insufficient information is available on
toxicologic effects or when the estimate
of the level of the pollutant in ambient
water based on organoleplic effects is
lower than Ihe level calculated from
toxicologic data. It should be recognized
that criteria based solely on
crvanokptic effects do net necessarily
represent approximations of acceptable
risk levels for human health.
  Several ambient water quality criteria
documents deal with classes of
compounds which include chemicals
exhibiting varying degrees of structural
similarity. Because prediction of
biological effects based solely on
structural parameters is difficult, the
derivation of compound-specific criteria
is preferable to'a class criterion. A
compound-specific criterion is defined
as a level derived from data on each
individual subject compound that does '
not represent  a significant risk to the
public. For some chemical classes,
however, a compound-specific criterion
cannot be derived for each member of a
class. In such Instances, it is sometimes
justifiable to derive a class criterion in
which available data on one member of
a class may be used to estimate criteria
for other chemicals of the class because
e sufficient data base is not available
for those  compounds.
  For some chemicals and chemical
classes, the data base was judged to be
insufficient for the derivation of a
criterion. In those cases, deficiencies in
the available Information are detailed.
///. Approach
  The human health effects chapters
attempt to summarize all information on
the individual chemicals or classes of
chemicals which might be useful in the
risk assessment process to develop
                     water quality criteria. Although primary
                     emphasis is placed on identifying
                     epidemiologic and toxicologic data,
                     these assessments typically contain
                     discussions on four topics: existing
                     levels of human exposure,
                     phannacoklnetics, toxic effects, and
                     criterion formulation.
                       For all documents, an attempt is made
                     to include the known relevant
                     information. Review articles and reports
                     are often used in the process of data
                     evaluation and synthesis. Scientific
                     Judgment is exercised in the review and
                     evaluation of the data in each document
                     and in the identification of the adverse
                     effects against which protective criteria
                     are sought. la addition, each of these  •
                     documents is reviewed by a peer
                     committee of scientists familiar with the
                     specific compound(s). These work
                     groups evaluate the quality of the
                     available data, the completeness of the
                     data summary, and the validity of the
                     derived criterion.
                       In the analysis and organization of the
                     data, an attempt is made to be
                     consistent with respect to the format
                     and the application of acceptable
                     scientific principles. Evaluation
                     procedures used in the hazard
                     assessment process follow the principles
                     outlined by the National Academy of
                     Sciences in Drinking Water and Health
                     (1077) and the guidelines of the
                     Carcinogen Assessment Group of the
                     U.S. EPA.
                     A. Exposure
                       The exposure section of the health
                     effects chapters reviews known
                     information on current levels of human
                     exposure to the individual pollutant
                     from all sources. Much of the data was
                     obtained from monitoring studies of air,
                     water, food, soil, and human or animal
                     tissue residues. The major purpose of
                     this section is to provide background
                     information on the contribution of water
                     exposure relative to all other sources.
                     Consequently, the exposure section
                     includes subsections reviewing different
                     routes of exposure including water and '
                     food ingestion, inhalation, and dermal
                     contact.
                       Information on exposure can be
                     valuable in developing and assessing a
                     water quality criterion. In these
                     documents exposure from consumption
                     of contaminated water and
                     contaminated fish and shellfish products
                     is used in criterion formulation. Data for
                     all modes of exposure are useful in
                     relating total intake to the expected
                     contribution from contaminated water.
                     fish, and shellfish. In addition,
                     information for all routes of exposure,
                     not limited to drinking water and fish
                     and shellfish ingestion, can be used to
 Justify or assess the feasibility of the
 formulation of criteria for ambient
 water.
   The use of fish consumption a's an
 exposure factor requires the
 quantitation of pollutant residues in the
 edible portions of the ingested species.
 Accordingly, bioconcentration factors
 (BCFs) are used to relate pollutant
 residues in aquatic organisms to the
 pollutant concentration in the ambient
 waters in which they reside.
   To estimate the average per capita
 intake of a pollutant due to consumption
 of contaminated fish and shellfish the
 results of a diet survey were analyzed to
 calculate the average consump tion of
' freshwater and-eetuarine fish and
 shellfish (U.S. EPA. 1980). A species is
 considered to be a consumed freshwater
 or estuarine fish and shellfish species if
 at some stage in its life cycle, it is
 harvested from fresh or estuarine water
 for human consumption in significant
 quantities (Stephan, 1980].
   Three different procedures are used to
 estimate the weighted average BCF
 depending upon the lipid solubility of
 the chemical and the availability of
 bioconcentration data.
   For lipid-soluble  compounds, the
 average BCF is calculated from the
 weighted average percent lipids in the
 edible portions of consumed freshwater
 and estuarine fish and shellfish which
 was calculated from data on
 consumption of each species and its
 corresponding percent lipids to be 3.0
 percent (Stephen, 1980). Because the
 steady-state BCFs for lipid-soluble
 compounds are proportional to percent
 lipids, bioconcentration factors for fish
 and shellfish can be adjusted to the
 average percent lipids for aquatic
 organisms consumed by Americans. For
 many lipid-soluble  pollutants,  there
 exists at least one BCF for which the
 percent lipid value was measured for the
 tissues for which the BCF is determined.
   With 3.0 percent  as the weighted
 average percent lipids for freshwater'
 and estuarine fish and shellfish in the
 average diet, a BCF, and a
 corresponding percent lipid value, the
 weighted average bioconcentration
 factor can be calculated.
   Example:
   Weighted average percent lipids for
     average diet=3.0 percent
   Measured BCF of 17 for
     trichloroethylene with bluegills at
     4.8 percent lipids
   Weighted average. BCF for average
     diet equals

         17 x  3.OX = 10.6
                4.8%

-------
                  Federal Register / Vol. 45. No. 231 / Friday. November 28, 1980 / Notices
                                                                      79349
  As an estimate, 10.6 Is used for the
    BCF.
  In those cases where an appropriate
bioconcentration factor Is not available,
the equation "Log BCF= (0.85 Log P)-
0.70" can be used (Veith, et al. 1979) to
estimate the BCF for aquatic organisms
containing about 7.6 percent lipids
(Veith, 1880) from.the octanol/water
partition coefficient P. An adjustment
for percent llplds in the average diet
versus 7,6 percent Is made in order to
derive the weighted average
bioconcentration factor.
  For non-lipid-sofuble compounds, the
available BCFs for the edible portion of
consumed freshwater and estuarine  fish
and shellfish are weighted according to
consumption factors to determine a
weighted BCF representative of the
average diet
B. Pharmacokinetjcs
  This section summarizes the available
information on the absorption,
distribution, metabolism, and
elimination of the compound(s) in
humans and experimental mammals.
Conceptually, such  Information is useful
in validation of inter- and intraspecies
extrapolations, and In characterizing the
modes of toxic action. Sufficient .
Information on absorption and excretion
in animals, together with a knowledge of
ambient concentrations in water, food,
and ah-, could be useful in estimating
body burdens of chemicals in the human
population. Distribution data which
suggest target organs or tissues are
desirable for interspecies comparison
techniques. In terms of the derivation of
criteria, pharmacoklnetic data are
essential to estimate equivalent oral *
doses based on data from Inhalation or
other routes of exposure.

C. Effects
  This section summarizes information
on biological effects in both humans  and
experimental mammals resulting in;
acute, subacute, and chronic toxiclty,
synergism and/or antagonism,
teratogenicity. mutagenidty. or
carcinogenicity;
  The major goal of this section Is to
survey the suitability of the data for use
in assessment of hazard and to
determine which biological end-point,
i.e., non-threshold, threshold, or
orgaholeptic, should be selected for use
in criterion formulation.
  Because this section attempts to
assess potential human health effects,
data on documented human effects are
thoroughly evaluated. However, several
factors inherent in human
epidemiological studies usually preclude
the use of such data in generating water
quality criteria. These problems, as
summarized by the National Academy
of Sciences (NAS, 1977) are as follows;
  1, Epidemiology cannot tell what
effects a material will have until after
humans have been exposed. One must
not conduct .what might be hazardous
experiments on man,
  2, If exposure has been ubiquitous, it
may be impossible to assess the effects
of a material, because there Is no
unexpoaed control group. Statistics of
morbidity obtained before use of a new
material can sometimes be useful, but
when latent periods are variable and
times of Introduction and removal of
materials overlap, historical data on
chronic effects are usually
unsatisfactory.
  3. it is usually difficult to determine
doses in human exposures.
  4, Usually, it is hard to identify small
changes in common effects, which may
nonetheless be important If the
population Is large.
  5. Interactions in a "nature-designed"
experiment usually cannot be
controlled.
  Although these problems often
prevent the use of epidemiological data
in quantitative risk assessments,
qualitative similarities or differences
between documented effects in humans
and observed effects in experimental
mammals are extremely useful In testing
the validity of animal-to-man
extrapolations. Consequently, in each
case, an attempt is made to identify and
utilize both epldemiologic and animal
dose-response data. Criteria derived
from such  a confirmed data base are
considered to be reliable. .
  The decision to establish a criterion
based on a non-threshold model is made
after evaluating all available
Information on carcinogenicity and
supportive Information on mutagenicity.
The approach and conditions for the
qualitative decision of carcinogenicity
are outlined in the U.S. EPA Interim
Cancer Guidelines {41FR 21402), in a
report by Albert, et al. (1977), and In the
Lnteragency Regulatory Liaison Group
(IRLG) guidelines on carcinogenic risks
(IRLG. 1979). It is assumed that a
substance which induces a statistically
significant carcinogenic response in
animals has the capacity to cause
cancer in humans. A chemical which
has not induced a significant cancer
response in humans or experimental
animals is not identified as a
carcinogen, even though its metabolites
or close structural analogues might
induce a carcinogenic response or it was
shown to be mutagenic in an in vitro
system.
  It is recognized that some potential
human carcinogens may not be
identified by the guidelines given above.
For example, compounds for which
there is plausible but weak qualitative
evidence of carcinogenicity in        *
experimental animal systems (such as
data from mouse skin painting or strain
A mouse pulmonary adenoma) would be
Included In this category. The derivation
of a criterion for human consumption
from these studies In not valid.
regardless of the qualitative outcome. In
addition, there are certain compounds -•
(e.g., nickel and beryllium) which were
shown to be carcinogenic in humans
after inhalation exposure by chemical
form, but have induced thus far no
response in animals or humans via
ingesting their soluble salts.
Nevertheless, a non-threshold criterion
is developed for beryllium because
tumors have been produced in animals
at a site removed from the site of
administration; in contrast, a threshold
criterion is recommended for nickel
because there is no evidence of tumors
at sites distant resulting from
administration of nickel solutions by
either Ingestion,or Injection.
  For those compounds which were not
reported to induce carcinogenic effects
or for those compounds for which
carcinogenic data are lacking or
Insufficient, an attempt is made to
estimate a no-effect level. In many
respects, the hazard evaluation from
these studies is similar to that of
bloassays for carcinogenldty. In order
to more closely approximate conditions
of human exposure, preference is given
to chronic studies involving oral
exposures in water or diet over a
significant portion of the animal life
span. Greatest confidence is placed In
those studies which demonstrate dose-
related adverse effects as well as no-
effect levels.
  There is considerable variability in
the biological endpoints used to define a
no-effect level. They may range from
gross; effects, such as mortality, to  more
subtle biochemical, physiological, or
pathological changes. Teratogenicity,
reproductive impairment, and
behavioral effects are significant toxic
consequences of environmental
contamination. In instances where
carcinogenic or other chronic effects
occur at exposure levels, below those
causing teratogenicity, reproductive
impairment,  or behavioral effects, the
former are used in deriving the criterion.
For most of the compounds evaluated
thus far, teratogenicity and reproductive
impairment occur at doses near
maximum tolerated levels with dose
administration schedules well above
estimated environmental exposure
levels. Moreover, information on
behavioral effects, which could be of

-------
70350
                   Federal Register / Vol. 45.  No. 201  /  Friday,  November 28, 1980  /  Notices
significance, is not available for most of
the compounds under study.
Consequently, most NOAELs derived
from chronic studies are based either on
gross toxic effects or on effects directly
related; to functional impairment or
defined pathological lesions.
  For, compounds on which adequate
chronic tosJcity studies,are not
available, studies on acute and subacute
toxicity assume greater significance.
Acute toxicity studies usually Involve
single exposures at lethal or near lethal
doses. Subacute studies often involve
exposures exceeding 10 percent of the
life span of the test organism, eug.. 90
days for the rat with an average life
span of 30 months. Such .studies are
useful in establishing the nature of the
compound's toxic effects and other
parameters of compound toxicity, such
as target organ effects,  metabolic1
behavior, physiological/biochemical
effects, and patterns of retention and
tissue distribution. The utility of acute
and subacute studies in deriving
environmentally meaningful NOELs is
uncertain, although McNamara [1076]
has developed application factors for
such derivations.
  In some cases where adequate data
are not available from studies, utilizing
oral routes of administration, no-effect
levels for oral exposures may be
estimated from dermal  or Inhalation
studies. Such estimates involve
approximations of the total  dose
administered based on  assumptions
about breathing rates and/or magnitude
of absorption.

D. Criterion Rationale
  This section reviews  existing
standards for the chemical(s),
summarizes data on current levels of
human exposure, attempts to identify
special groups at risk, and defines the
basis for the recommended  criterion,
  Information on existing standards is
included primarily for comparison with
the proposed water quality criteria.
Some of the present standards, such as
those recommended by the
Occupational Safety and Health
Administration (OSHA) or the American
Conference of Governmental Industrial
Hygienists (ACG1H), are based on
toxicologic data but are intended as
acceptable levels for occupational
rather than environmental exposure.
Other levels, such as those
recommended by the National Academy
of Sciences in Drinking Water and
Health (1E77) or in the U.S. EPA Interim
Primary Drinking Water Standards, are
more closely related to  proposed water
quality criteria. Emphasis is placed on
detailing the basis for the existing
standards wherever possible.
  Summaries of current levels of human
exposure, presented in this section,
specifically address the suitability of the
data to derive water quality criteria. The
Identification of special groups at risk,
either because of geographical or
occupational differences in exposure or
biological differences in susceptibility to
the compoundfs], focuses on the impact
that these groups should have on the
development of water quality criteria.
  The basis for the recommended
criteria section summarizes and
qualifies all of the data used in
developing the criteria.             •
IV. Guidelines for Criteria Derivation
  The derivation of water quality
criteria from laboratory animal toxicity
data is essentially a two-step procedure.
First, a total daily intake for humans
must be estimated which establishes
either a defined level of risk for non-
threshold effects or a no-effect level for
threshold effects. Secondly, assumptions
must be,made about the contribution of
contaminated water and the
consumption offish/shellfish to the total
daily intake of the chemical. These
estimates are then used to establish the
tolerable daily intake and consequently
the water quality criterion,
A, Non-Threshold Effects
  After the decision has been made that
a compound has the potential for
causing cancers in humans and that  .
data exist which permit the derivation
of a criterion, the water concentration
which is estimated to cause a lifetime
carcinogenic risk of 10"* is determined.
The lifetime carcinogenidty risk Is the
probability that a person would get
cancer sometime in his or her life
assuming continuous exposure to the
compound. The water concentration is
calculated by using the low-dose
extrapolation procedure proposed by
Crump (t980). This procedure is an
improvement on the multistage low dose
extrapolation procedure by Crump, et al.
(1977).
  The data used for quantitative
estimates are of two types: (1) lifetime
animal studies, and (2) human studies
where excess cancer risk has been
associated with exposure to the agent.
In animal studies it is assumed, unless
evidence exists to the contrary, that if a
carcinogenic response occurs at the
dose levels used in the study, then
proportionately lower responses will
also occur at all lower doses, with an -
incidence determined by the
extrapolation model discussed below.
  1. Choice of Model.
•  There is no really solid scientific basis
for any mathematical extrapolation
model which relates carcinogen
exposure to cancer risks at the
extremclylbtv levels of concentration
that must be dealt with in evaluating the
environmental hazards. For practical
reasons, such low "levels of risk cannot
be measured directly either.using animal
experiments pr.epidemiologic studies.
We must, therefore, depend on  our
current understanding of the
mechanisms of carcinogenesis for
guidance as to which risk model to use.
At the present time, the dominant view
of the carcinogenic process involves the
concept that most agents which cause
cancer also cause irreversible damage to
DN A. This position is reflected by the
fact that a very large proportion of
agents which cause cancer are  also
mutajenic. There is reason to expect
that the quontal type of biological
response that is characteristic of
mutagenesls is associated with a linear
non-threshold dose-response
relationship. Indeed, there is  substantial
evidence frcm mutagenesia studies with
both ionizing radiation and with a wide
variety of chemicals that this type of
dose-response model is the appropriate
one to use. This is particularly true at
the lower end of the dose-response
curve; at higher doses, there can be an
upward curvature, probably reflecting
the effects of multistage processes on
the mutagenic response. The  linear non-
threshold dose-response relationship is
also consistent with the relatively few
epldemlological studies of cancer
responses to specific agents that contain
enough information to make the
evaluation possible (e.g., radiation-
induced leukemia, breast and thyroid
cancer, skin cancer induced by arsenic
in drinking water, and liver cancer
induced by aflatoxin in the diet). There
is also some evidence  from animal
experiments that is consistent with the
linear non-threshold hypothesis (e.g.,
liver tumors induced in mice  by 2-
acetylamlnofluorene in the large scale
ED»i study at the National Center of
Toxicologies! Research, and  the
initiation stage of the two-stage
carcinogenesis model in the rat liver and
the mouse skin],
  Because it has the best, albeit limited.
scientific basis of any of the current
mathematical extrapolation models, the
linear non-threshold model has been
adopted as the primary .basis for risk
extrapolation to low levels of the dose-
response relationship. The risk
assessments made with this model
should be regarded as conservative,
representing the most plausible upper
limit for the risk; ijg., the,true risk is not
likely to be higher than the estimate, but
it could be smaller.

-------
                   Federal Register / Vol.  45. No. 231 / Friday. November 28. 1980  /  Notices
                                                                                                             79351
  The mathematical formulation chosen
 to describe the linear, non-threshold
 dose-response relationship at low doses
 islUie improved multistage model
 developed by Crump (1980). This model
 employs enough arbitrary constants to
"be able to fit almost any monotonically
 increasing dose-response data and it
 incorporates a procedure for estimating
 the largest possible linear slope (in the
 05 percent confidence limit sense) at low
 extrapolated doses that is consistent
 with the data at all dose levels of the
 experiment. For this reason, it may be
 called a "linearized" multistage model.
  2. Procedure of Low-Dose
 Extrapolation Based on Animal
 Carcinogenicity Data.
  A. Description of the Extrapolation
 Model
  Let P(d] represent the lifetime risk
 (probability) of cancer at dose d. The
 multistage model has the form
 Pfd}=l-exp[-(q.+q,d+q,ds+. . . +qkdk]]
 where:
  q,>0,andl=0,l. 2,. .
Equivalently,
  A(d)=l-exp [-
where:
                    . ,k
                           .  . +qkd"J]
      A(d)  -  P(d)*P(o).
                  1  -  P(o)

is the extra risk over background rate at
dosed.
  The point estimate of the coefficients
qi, 1=0, 1, 2 ..... k. and consequently
the extra risk function A(d) at any given
dose d, is calculated by maximizing the
likelihood function of the data.
  The point estimate and the 95 percent
upper confidence limit of the extra risk
A(d) are calculated by using the
computer program GLOBAL 79
developed by Crump and Watson (1979).
Upper 96 percent confidence limits on
the extra risk and lower 95 percent
confidence limits on the dose producing
a given risk are determined from a 95
percent upper confidence limit, q/. on
parameter qi. Whenever qi-*0, at low
doses extra risk A(d) has approximately
the form A(d)=q,xd. Therefore. qiXd
is a 95 percent upper confidence limit on
the extra risk and R/qi* is a 95 percent
lower confidence limit on the dose
producing an extra risk of R. Let L» be
the maximum value of the log-likelihood
function. The upper limit qt* is
calculated buy increasing q1 to a value
q,* such that when the log-likelihood is
again maximized subject to this fixed
value q/ for the linear coefficient, the
resulting maximum value of the log-
likelihood Li satisfies the equation
2(L,-L,) = 2.70554
where 2.70554 is the cumulative 90
percent point of the chi-square
distribution with one degree of freedom.
which corresponds to a 95 percent upper
limit (one-sided). This approach of
computing the upper confidence limit for
the extra risk A(d) is an improvement on
the Crump, et al. (1977) model. The
upper confidence limit for the extra risk
calculated at low doses is always linear.
This is conceptually consistent with the
linear nonthreshold concept discussed
earlier. The slope qi* is taken as an
upper bound of the potency of the
chemical in Inducing cancer at low
doses.
  In fitting the  dose-response model the
number of terms in the polynomial g is
chosen equal to (h-1), where h is the
number of dose groups in the
experiment, including the control group.
  Whenever the multistage model does
not fit the data sufficiently, data at the
highest dose is deleted and the model is
refitted to the rest of the data. This is
continued until an acceptable fit to the
data is obtained. To determine whether
or not a fit is acceptable, the chi-square
statistic:
                                                             (1  - Pi)
                                       is calculated, where N, is the number of
                                       animals in the 1th dose group, X( is the
                                       number of animals in the 1th dose group
                                       with a tumor response; P, is the
                                       probability of a response in the 1th dose
                                       group estimated by fitting the multistage
                                       model to the data, and h is the number
                                       of remaining groups.
                                         The fit is determined to be
                                       unacceptable whenever chi-square (X1)
                                       is larger than the cumulative 99 percent
                                       point of the chi-square distribution with
                                       f degrees of freedom, where f equals the
                                       number of dose groups minus the
                                       number of non-zero multistage
                                       coefficients.
                                         3. Selection and Form of Data used to
                                       Estimate Parameters in the
                                       Extrapolation Model.
                                         For some chemicals, several studies in
                                       different animal species, strains, and
                                       sexes each conducted at several doses
                                       and different routes of exposure are
                                       available. A choice must be made as to
                                       which of the data sets from several
                                       studies are to be used in the model. It is
                                       also necessary to correct for metabolism
                                       differences between species and for
                                       differences in absorption via different
                                       routes of administration.  The
                                       procedures, listed below, used in
                                       evaluating these data are consistent
                                       with the estimate of a maximum-likely-
                                       risk.
  a. The rumor incidence data are
separated according to organ sites or
tumor types. The set data (i.e.. dose and.
tumor incidence) used in the model is
set where the incidence is statistically
significantly higher than the control for
at least one test dose level and/or
where the tumor incidence rate shows a
statistically significant trend with
respect to dose level. The data set which
gives the highest estimate of lifetime
carcinogenic risk q/  is selected in most
cases. However, efforts are made to
exclude data sets which produce
spuriously high risk estimates because
of a small number of animals. That is, if
two sets of data show a similar dose-
response relationship and one has a
very small sample size,'the set of data
which, has the larger sample size is
selected for calculating the carcinogenic
potency,
  b. If there are two or more data sets of
comparable size which are identical
with respect to species, strain, sex, and
tumor sites, the geometric mean of q,*,
estimated from each of these data sets is
used for risk assessment. The geometric
mean of numbers A,,  A,	Am is
defined as (AiXA,X  . . .  XAn,)1/m
  c. If sufficient data  exist for two or
more significant tumor sites in the same
study, the number of animals with at
least one of the specific tumor sites
under consideration is used as Incidence
data in the model.
  d. Following the suggestion of Mantel
and Schneidennan (1975), we assume
that rag/surface area/day is an
equivalent dose between species. Since
to a close approximation the surface
area is proportional to the %rds power
of the weight as would be the case for a
perfect sphere, the exposure in mg/%rds
power of the body weight/day is
similarly considered to be an equivalent
exposure. In an animal experiment, this
equivalent dose is computed in the
following manner
Let:
Induration of experiment
U=duratlon of exposure
m=average dose per day In mg during
    administration of the agent (I.e.. during 1,)
VV=average weight of the experimental
    animal.
Then, the lifetime average exposure  is

          .        le  X HI
                                                     Le x W2/3

                                         Often exposures are not given in units
                                       of rag/day, and it becomes necessary to
                                       convert the given exposures into mg/
                                       day. For example, in most feeding
                                       studies, exposure is expressed as ppm in
                                       the diet. In this case the exposure (mg/
                                       day] is  derived by: m = ppm  x F x r

-------
 79352
       Federal  Register / Vol. 45. No. 231 / Friday. November 28. 1980 / Notices
 where ppm is parts per million of the
 carcinogenic agent in the diet. F is the
 weight of the food consumed per day in
 kgms, and r is the absorption fraction.
   In the absence of any data to the
 contrary, r is assumed to be one. For a
 uniform diet the weight of the food
 consumed is proportional to the calories
 required, which, in turn, is proportional .
 to the surface area or the %rds power of
 the weight, so that; mappmxW*'*xr or
    m
ppm
   ,2/3
 rW
  As a result, ppm in the diet is often
 assumed to be an equivalent exposure
 between species. However, we feel that
 this is not justified since the calories/kg
 of food is significantly different in the
 diet of man vs. laboratory animals,
 primarily due to moisture content
 differences. Instead, we use an
 empirically derived food factor, f=F/W,
 which is the fraction of a species body
 weight that Is consumed per day as
 food. We use the rates given below.
                              w
                                   i
Mice.
                             70   04)28
                              0-35 0.05
                              0.03 0,13
Thus, when the exposure is given as a
certain dietary concentration in ppm, the
exposure in ing/ W2' * is
     ITL     a  ppm x  F  «.
 r  x  W2/3      W2/3
               -  PP«n x  f  x wl/3
  When exposure is given in terns of
mg/kg/day<=m/Wr=s the conversion is
simply:

   •>..   a  S X Wl/3
 rWZ/3

  When exposure is via inhalation, the
calculation of dose can be considered •
for two cases where (1) the carcinogenic
agent is either a completely water-
soluble gas or an aerosol and is
absorbed proportionally to the amount
of air breathed in, and (2) where the
carcinogen is a poorly water-soluble gas
which reaches an equilibrium between
the air breathed and the body
compartments. After equilibrium is
reached, the rate of absorption of these
agents is expected to be proportional to
metabolic rate, which in turn is
proportional to the rate of oxygen
consumption, which in turn is a function
of surface area.
                                       Case 1
                                         Agents that are in the form of
                                       paniculate matter or virtually
                                       completely absorbed gases such as SO»
                                       can reasonably be expected to be
                                       absorbed proportional to the breathing
                                       rate. In this case the exposure in mg/day
                                       may be expressed as: m=Ixvxr where
                                       I is inhalation rate per day in m', v is
                                       mg/m* of the agent in air, and r is the
                                       absorption fraction.
                                         The Inhalation rates, I, for various
                                       species can be calculated from the
                                       observation [FASEB, 1074) that 25 gm
                                       mice breathe 34.B liters/day and 113 gm
                                       rats breathe 105 liters/day. For mice and
                                       rats of other weights, W, (expressed in
                                       kg], the surface area proportionality can
                                       be used to determine breathing rates (in
                                       m'/day) as follows:
                                         For mice, 1=0.0345 (W/0.025)*'»mV
                                       day
                                         For rats, 1-0.105 (W/0.113)"tayday
                                         For humans, the values of 20 m'/day *
                                       is adopted as a standard breathing rate
                                       (ICRP, 1977).
                                         The equivalent exposure in mg/W*'*
                                       for these agents can be derived from the
                                       air intake data in a way analogous to
                                       the food intake data. The empirical
                                       factors for the air intake per kg per day,
                                       i=I/W based upon the previously stated
                                       relationships, are as tabulated below.
                                                                     w
                                                                         w
                                       Men	
                                       Rat	
                                       Mica	
                                                        70    0.28
                                                       ,  OM  0.04
                                                       .  0.03  U
                           Therefore, for participates or completely
                           absorbed gases, the equivalent exposure
                           iamg/Wtr>is;
                             m  ......
                            w2/3
                           In the absence of empirical data or a
                           sound theoretical argument to the
                           contrary, the fraction absorbed, r, is
                           assumed to be the same for all species.
                           Case 2
                             The dose in mg/day of partially
                           soluble vapors is proportional to the 0,
                           consumption which in turn is
                           proportional to Wa"and to the
                           solubility of gas in body fluids, which
                           can be expressed as an absorption
                           coefficient r for the gas. Therefore, when
                           expressing the Ot consumption as 0«=k
                           W" 8, where k is a constant independent
                             * From "Recommendation of the International
                           Commission on Radiological Protection," page 9, the
                           average breathing rate IB 10* cm'per 8-hour work
                           day and 2 X10' cm' in 24 hours.
                                                                  of species, it follows that m=k W"»
                                                                  x r or

                                                                                 kvr
As with Case 1, in the absence of
experimental information or a sound
theoretical argument to the contrary, the
absorption fraction, r, is assumed to be
the same for all species. Therefore, for
these substances a certain concentration
in ppm or ji/ma In experimental animals
is equivalent to the same concentration
in humans. This is supported by the
observation that the minimum alveolar
concentration, necessary to produce a
given "stage" of anesthesia, is similar in
man and animals (Dripps, et al. 1977).
When the animals were exposed via the
oral route and human exposure is via
inhalation or vice-versa, the assumption
is made, unless there is pharmacokinetic
evidence to the contrary, that absorption
is equal by either exposure route.
  e. If the duration of experiment (L.) is
leas than the natural life span of the test
animal (L), the slope q«*. or more
generally the exponent g(d), is increased
by multiplying a factor (L/L,)1. We
assume that if the average dose, d, is
continued, the age specific rate of
cancer will continue to increase as a
constant function of the background
rate. The age specific rates for humans
increase at least by the 2nd power of the
age and often by a considerably higher
power, aa demonstrated by Doll (1071).
Thus, we would expect the cumulative
tumor rate to increase by at least the 3rd
power of age. Using this fact, we assume
that the slope qi*. or more generally, the
exponent g(d), would also increase by at
least the 3rd power of age. As a result, if
the slope 0,1* [or g(d)] is calculated at
age L,. we would expect that if the
experiment had been continued for the
full life span, L, at the given average
exposure, the slope qt* [or g[d)] would
have been increased by at least (L/L*)1.
  This adjustment is conceptually
consistent to the proportional hazard
model proposed by Cox (1972) and the
time-to-tumor model considered by
Crump, et al. (1977) where the
probability of cancer at age t and dose d
is given by P(d,t)=l-exp[-f(t)Xg(d)]
  4. Calculation of Carcinogenic Potency
Based on Human Data. If human
epidemiology studies and  sufficiently
valid exposure information are available
for the compound, they are always used
in some way. If they show a
carcinogenic effect, the data are
analyzed to give an estimate of the
linear dependence of cancer rates on
lifetime average dose, which is
equivalent to the factor q,*. If they show

-------
                   Federal Register  /  Vol. 45. No. 231  /  Friday.  November 28. 1980 / Notices     	79353
 no carcinogenic effect when positive
 animal evidence la available, then It is
 assumed that a risk does exist but it Is
 smaller than could have been observed
 in the epidemiologic study, and an upper
 limit of the caacer Incidence is
 cajculated assuming hypotheticaJly that
 the true incidence is just below the level
 of detection in the cohort studied, which
 is determined largely by the cohort size.
 Whenever possible, human data are
 used in perference to animal bioassay
 data.
   In human studies, the response is
 measured in terms of the relative risk of
 the exposed cohort of individuals
 compared to the control group.in the
 analysis of this data, it is assumed that
 the excess risk, or relative risk minus
 one, R(X)-1. is proportional to the
 lifetime average exposure, X, and that it
 is the same for all ages. It follows that
 the carcinogenic potency Is equal to
 PUX)-1]/X multiplied by the.lifettme
 risk at that site in the general
 population. Except for an unusually
 well-documented human study, the
 confidence limit for the excess risk is
 not calculated, due to the difficulty in
 accounting for the uncertainty inherent
 in the data (exposure and cancer
 response).
  5. Calculation;of Water Quality
 Criteria/After the value of q/ in (mg/-
 kg/day}"1 has been determined, the
 lifetime risk, P, from an average daily
 exposure of x mg/kg/day is found from
 the equation P=qi*x Therefore, if the
 lifetime risk is set at P=.l
-------
79354
Federal  Register / Vol. 45. No. 231  /  Friday. November 28. 1980 /  Notices
Indication of carclnogenieJty" la
interpreted as the absence of
carcinogenic! ty data from animal
experimental studies or human
epidemiology. Available short-term
carcinogeniclty screening tests are
reported in the criteria documents, but
they are not used either for derivation of
numerical criteria nor to rule out the
uncertainty factor approach.
  Because of the high degree of
judgment involved in the selection of a
safety factor, the criterion derivation
section of each document should
provide a detailed discussion and
justification far both the selection of the
safety factor and the data to which it ia
applied. This discussion should reflect a
critical review of the available data
base. Factors to be considered include
number of animals, species, and
parameters tested; quality of controls;
doae levels; route; and dosing schedules.
An effort should be made to
differentiate between results which
constitute a lexicologically sufficient
data base and data which may be
spurious In nature.
  2. Use of Acceptable Daily Intake
(ADI). For carcinogens, the assumption
of low dose linearity precludes the
necessity for defining total exposure in
the estimation of increased incremental
risk. For non-carcinogens, ADIs and
criteria derived therefrom are calculated
from total exposure data that Include
contributions from the diet and air. The
equation used to derive the criterion (C)
is: C=ADI-(DT+IN}/[2 l-f-(Q,OQ8! kg
X R)] where 21 is assumed daily water
consumption, 0.0085 kg is assumed daily
fish consumption, R is bioconcentration
factor in units of I/kg, DT is estimated
non-fish dietary intake, and IN is
estimated daily intake by inhalation.
  If estimates of IN and DT cannot be .
provided from experimental data, an
assumption must be made concerning
total exposure. It Is recognized that
either the inability to estimate DT and
IN due to lack of data or the wide
variability in DT and IN  in different
states may add an additional element of
uncertainty to the criterion formulation
process. In terms of scientific validity,
the accurate estimate of the Acceptable
Daily Intake is the major factor in
satisfactory derivation of water quality
criteria.
  3. Use of Threshold Limit Values or
Animal Inhalation Studies. Threshold
Limit Values (TLVs) are  established by
the American Conference of
Governmental and Industrial Hygienists
(ACGIH) and represent 8-hour time-
weighted average concentrations in air
that are intended to protect -workers
from various adverse health effects over
a normal working lifetime. Similar
                     values are set by NIOSH (criteria) and
                     OSHA (standards] for 10- and 8-hour
                     exposures, respectively. To the extent
                     that these values are baaed on sound
                     toxicologic assessments and have been
                     protective in the work environment, they
                     provide useful information for deriving
                     or evaluating water quality criteria.
                     However, each TLV must be carefully
                     examined to determine if the basis of
                     the TLV contains data which can be
                     used directly to derive a water quality
                     criterion using the uncertainty factor
                     approach. In addition, the history of
                     each TLV must be examined to assess
                     the extent to which it has assured
                     worker safety. In each case, the types of
                     effects against which TLVs are designed
                     to protect are examined In terms of their
                     relevance to exposure from water. It
                     must be demonstrated that the chemical
                     is not a localized irritant and that there
                     is no significant effect at the site of
                     entry irrespective of the routes of
                     exposure (i.e., oral or inhalation).
                       If the TLV or similar value is
                     recommended as the basis of the
                     criterion, consideration of the above
                     points is explicitly stated in the  criterion
                     derivation section of the document.
                     Particular emphasis is placed on the
                     quality of the TLV relative to the
                     available toxicity data that normally is
                     given priority over TLVs or similar
                     established values. If the TLV can be
                     Justified as the basis for the clrterion,
                     then the problems associated with the
                     estimation of acceptable oral doses from
                     inhalation data must be addressed.
                       Estimating equivalencies of dose-
                     response relationships from one route of
                     exposure to another introduces  an
                     additional element of uncertainty  in the
                     derivation of criteria. Consequently,
                     whenever possible, ambient water
                     quality criteria should be based on data
                     involving oral exposures, if oral data are
                     insufficient, data from other routes of
                     exposure may be useful in the criterion
                     derivation process.
                       Inhalation data, including TLVs or
                     similar values, are the most common
                     alternatives to oral  data. Estimates of
                     equivalent doses can be baaed upon: (1)
                     available pharmacokinetic data for oral
                     and inhalation routes, (2) measurements
                     of absorption efficiency from ingested or
                     inhaled chemicals, or (3) comparative
                     excretion data when the associated
                     metabolic pathways are equivalent to
                     those following oral ingestion or
                     Inhalation. Given that sufficient
                     pharmacokinetic data are available, the
                     use of accepted pharmacokinetic models
                     provides the most satisfactory approach
                     for dose conversions. However, if
                     available pharmacokinetic data are
                     marginal or of questionable quality,
pharmacokinetic modeling is
inappropriate.
  The Stokinger and Woodward (1958)
approach, or similar models based on
assumptions of breathing rate and
absorption efficiency, represents
possible alternatives when data are not
sufficient to justify pharmacokinetic
modeling. Such alternative approaches,
however, provide less satisfactory
approximations because they are not
based on pharmacokinetic data.
Consequently, in using the Stokinger
and Woodward or related models, die
uncertainties inherent In each of the
assumptions and the basis of each
assumption must be clearly stated in the
derivation of the criterion.
  The use of data pertaining to other
routes of exposure to derive water
quality criteria may also be considered.
As with inhalation data, an attempt is
made to use accepted toxicologic and
pharmacokinetic principles to estimate
equivalent oral doses. If simplifying
assumptions are used, their bases and
limitations must be clearly specified.
  Because of the uncertainties involved
in extrapolating from one route of
exposure to another and the consequent
limitations that this may place on the
derived criterion, the decision to
disallow such extrapolation and
recommend no criterion is highly
judgmental and must be made on a c.v^
by-case basis. A decision for or against
criteria derivation must balance the
quantity and quality of the available
data against a perceived risk to the
human population.
  If the Stokinger and Woodward (1S58)
approach is used to calculate an ADI
from a TLV, the general equation is:
ADI=TLVxBRxpExdxAA/(AoXSF)
where:
    s Acceptable daily intake in mg
TLVt= Concentration in air in mg/m'
DE*s Duration of exposure In hours per day
d =5 days/7 days
AA=Efficiency of absorption from air
AO= Efficiency of absorption bom oral
    exposure
SF=Safety factor following guidelines given
    above
BR= Amount of air breathed per day; assume
    10 IB*

For deriving an ADI from animal
toxicity data, the equation is:
ADI=eAxDBxdxAAXBRx70 kg/
(BWAxAoXSF) where:
ADI = Acceptable dally intake in mg
CA = Concentration in sir in mg/m*
Dt= Duration of exposure In hours per day
d= Number of days exposed/number of days
    observed
AA=Efficiency of absorption from air
BR = Volume of air breathed per day in m =
70 kg= Assumed human body weight
BWA=Body weight of experiments! animals
    in kg

-------
                    Federal  Register / Vol. 4S. No.  231 / Friday.. November 2B. 19BO  / Notices
                                                                         7B355
  0 = Efficiency of absorption from oral
     exposure
  ?m Safety factor following guidelines given
     above.
  lore formal pharmacokinetic models
  tust be developed on a compound-by-
  nmpound basis.
   It should be noted that the safety
  ictors used in the above formulae are
  Ucnded to account for species
 . ariability. Consequently, the mg/
  urface area/day conversion factor is
  ot used In the derivation of toxicity
  a sed criterion.

  *. Organoleptic Criteria
   Organoleptic criteria define
  oncentratlons of materials which
  npart undesirable taste and/or odor to
  /ater. In developing and utilizing such
  riteria two factors must be appreciated:
  le limitations of most organoleptic data
  rid the human health significance of
  rganoleptlc properties.
  The publications which report taste
  nd odor thresholds are, with very few .
  xceptions, cryptic in their descriptions
  'f test methodologies, number of
  ubjects tested, concentration: response
  elationships, and sensory
  .haracterisUca at specific
  .oncentrations above threshold. Thus,
  he quality of organoleptic data is often
  ignificantly less than that of toxicologlc
  iata used in establishing other criteria,
  Consequently, a critical evaluation of
  he available organoleptic data must be
 nade and the selection of the most
 ippropriate data base for the criterion
 nust be based on sound scientific
  udgment.             .
  Organoleptic criteria,are not based on
 oxicologic information and have no
 lirect relationship to potential  adverse
 lunian health effects. Although
 .ufficlently Intense organoleptic
 -.haracteristics could result in depressed
 luid Intake which, in turn, might
 iggravate a variety of functional disease
 itales (i.e., kidney and circulatory
 iiseases), such effects are not used In
  he derivation process of organoleptic
. -.riteria unless available data would
 ndicate an indirect human health effect
 /la decreased fluid consumption,
criteria derived solely from organoleptic
 iata are based upon aesthetic qualities
 inly.
  Since organoleptic and human health
 iffects criteria are based on different
 •ndpoints, a distinction must be made
 Between these two sets of information.
 in criteria summaries involving both
 types of data, the following format is_
 jscd:
  For comparison purposes, two approaches
 ^/ere used to derive criterion levels for
 	. Based on available toxicity data.
 •'or the protection ofT>ubIic health the derived
level is —-. Using available organolepttc
data, for controlling undesirable taste and
odor quality of ambient water the estimated
level is	. It should be recognized that
organoleptic data as a basis for establishing a
water quality criteria have no demonstrated
relationship to potential adverse human
health affects.
  In those instances where a level to
limit toxicity cannot be derived, the
following statement is to be
appropriately Inserted:
  Sufficient data are not available tor
•	to derive a level which would
protect against the potential toxicity of this .
compound.
D. Criteria for Chemical Classes
  A chemical class is broadly defined as
any group of chemical compounds which
are reviewed in a single risk assessment
document. In criterion derivation,
isomers should be regarded as a part of
a chemical class rather than as a single
compound. A class criterion is an
estimate of risk/safety which applies to
more than one member of a class. It
involves .the use of available data on
one or more chemicals of a class to
derive criteria for other compounds of
the same class In the event that there
are insufficient data  available to derive
compound-specific criteria.
  A class criterion usually applies to
each member of a class rather than to
the sum of the compounds within the
clasfl, While the potential hazards of
multiple toxicant exposure are not to be
minimized, a criterion, by definition,
most often'applies to an individual
compound; Exceptions may be made for
complex mixtures which are produced,
released, and toxicologieslly tested as
mixtures (e.g., toxaphene andPCBs). For
such exceptions, some attempt is made
to assess the effects of environmental
partitioning (I.e., different patterns of
environmental transport and
degradation) on the validity of the
criterion. If these effects cannot be
assessed, an appropriate statement of
uncertainty should accompany the
criterion..   .'.-   ; ; .••;•:    -..•.  ••, '•'  /
  Since relatively minor structural
changes within a class of compounds
can have pronounced effects on their
biological activities, reliance on class
criteria should be minimized. Whenever
sufficient toxicologic, data are available
on a chemical within a class, a
compound-specific criterion should be
derived. Nonetheless, for some chemical
classes, scientific judgment may suggest
a sufficient degree of similarity among
chemicals within a class to justify a
class criterion applicable to some of  all
members of a class.
  The development of a class criterion
takes into consideration the following:
  3, A detailed review of the chamica] and
physical properties of chemicals within the
group should be made. A dose relationship
within the class with respect to chemical
acitivity would suggest a similar potential to
reach common biological sites within tissues.
Likewise, similar lipld solubilities would
suggest the possibility of comparable
absorption and tissue distribution.
  2. Qualitative and quantitative data for
chemicals within the group are examined.
Adequate toxicologic data on a number of
compounds within a group provides a more
reasonable basis for extrapolation to other
chemicals of the same class than minimal
data on one chemical or a few chemicals
within the group.
  3. Similarities In the nature of the
toxicologic response to chemicals In the class
provides additional support for the prediction
that the response to other members of the
class may be similar. In contrast, where the
biological response has been shown to differ
markedly on a qualitative and quantitative
basis for chemicals within a class, the
extrapolation of a criterion to other members
of that class is not appropriate.
  4, Additional support for the validity of
extrapolation of a criterion to other members
of a class could be provided by evidence of
similar metabolic and pharmacokinetic data
for some members of the class. .
  Based on the above considerations, it
may be reasonable In some cases to
divide a chemical class into various
subclasses. Such divisions  could be
based on biological endpoints (e.g..
carcinogens /non-carcinogens), potency,
and/or sufficiency of data (e.g., a
criterion for some members of a class
but no criterion for others]. While no a
priori limits can be placed on the extent
of subclassification, each
subclaasification must be explicitly
justified by the available data.
  Class criteria, if properly derived and
supported, can constitute valid scientific
assessments of potential risk/safety.
Conversely, the development of a class
criterion from an insufficient data base
can lead to serious errors in
underestimating or overestimating risk/
safety and should be rigorously avoided.
Although scientific judgment has a
proper role in  the development of class
criteria, such criteria are useful and
defensible only if they are based on
adequate data and scientific reasoning.
The definition of sufficient data on
similarities in physical, chemical,
pharmacokinetic, or toxicologic
properties to justify a class criterion
may vary markedly depending on the
degree of structural similarity and the
gravity of the perceived risk.
Consequently, it is imperative that the
criterion derivation section of each
document in which a class criterion is
recommended explicity address each of
the key issues discussed above, and
define, as clearly as possible,  the

-------
  79356
Federal Register / Vol. 45. No. 231 / Friday.  November 28. 1980  / Notices
. limitations of the proposed criterion as
 well as the type of data needed to
 generate a compound-specific criterion.
   A class criterion should be abandoned
 when there is sufficient data availabe to
 derive a compound-specific criterion
 which protects against the biological
 effect of primary concern; e.g., the
 availability of a good subchronic study
 would not necessarily result in the
 abandonment of a class criterion based
 on potential carcinogenicity.
   The inability to derive a valid class
 criterion does not, and should not,
 preclude regulation of a compound or
 group of compounds based on.concern
 for potential: human health effects. The
 failure to recommend a criterion is
 simply a statement that the degree of
 concern cannot be  quantified based on
 the available data and risk assessment
 methodology.
 E. Essential Elements
   Some chemicals, particularly certain
 metals, are essential to biological
 organisms at low levels but may be
 toxic and/ or carcinogenic at high levels.
 Because of potential toxic effects, it is
 legitimate to establish criteria for such
 essential elements. However, criteria
 must consjder essentiality'and cannot
 be established at levels which would
 result In deficiency of the element in the
 human population.
   Elements are accepted as essential if
 listed by NAS Food and Nutrition Board
 or a comparably qualified panel.
 Elements not yet determined to be
 essential but for which supportive data
 on essentiality exists need to be  further
 reviewed by such a panel
   To modify the toxicity and
 carcinogenicity based  criteria,
 essentiality must be quantified either as
 a "recommended daily allowance"
 (RDAj or "minimum daily requirement"
 (MDR). These levels,are then compared
 to estimated daily doses associated  with
 the adverse effect of primary concern.
 The difference between the RDA or
 MDR and the daily doses causing a
 specified risk level  for carcinogens or
 ADIs for non-carcinogens defines the
 spread of daily doses from which the
 triterion may be derived. Because errors
 are inherent in defining both essential
 and maximum tolerable levels, the
 criterion is derived from dose levels  .
 near the center of such a dose range.
 The decision to use either the MDR or
 RDA is guided by the spread of the
 doses and the quality of the essentiality
 and toxicity estimates.
   The modification of criteria by
 consideration of essentiality must take
 into account all routes of exposure. If
 water is a significant source of the MDR
 or RDA, the criterion must allow for
                      attainment of essential intake.
                      Conversely, even when essentiality may
                      be attained from nonwater sources.
                      standard criteria derivation methods
                      may be adjusted if the derived criterion
                      represents a small fraction of the ADI or
                      MDR. On a case-by-case basis,'the
                      modification in the use of the guidelines
                      may include the use of different safety
                      factors for non-carcinogens or other
                      modifications which can be explicitly
                      justified.

                      F. Use of Existing Standards

                        For some chemicals for which criteria
                      are to be established; drinking water
                      standards already exist. These
                      standards represent not only a critical
                      assessment of literature, but also a body
                      of human experience since their
                      promulgation. Therefore, it is valid to
                      accept the existing standard unless
                      there is compelling evidence to the
                      contrary. This decision should be made
                      after considering the existing standards
                      vs. new scientific evidence which has
                      accumulated since the standards have
                      been established. There are several
                      instances where the peer review process
                      recommended usage of the present
                      drinking water standards.

                      Bibliography
                       Albert, RJE.. et aL 1977. .Rationale
                      developed by the Environmental Protection
                      Agency for the assessment of carcinogenic
                      risks. Jour. Natl. Can. Inst 58:1537.
                       Cox, C.R. 1972. Regression model and life
                      tables. Jour. Roy. Slat Soc. B. 34:187.
                       Crump, K.S; 1979. Dose-response problems
                      in carcinogenesis. Biometrics. 35:157.
                       Crump. K.S. 1980. An improved procedure
                      for low-dose carcinogenic risk assessment
                      from animal data. Jour. Environ. Path.
                      Toxicol. (In press].
                       Crump. K.S. and WAV. Watson. 1979.
                      GLOBAL 79. A FORTRAN program to
                      extrapolate dichotomous animal  '
                      carcinogenicity data to low dose. Natl. Inst.
                      Environ, Health Sci. Contract No. l-ES-2123.
                       Crump, K.S., et al. 1977. Confidence   •
                      intervals and test of hypotheses concerning
                      dose-response relations inferred from animal
                      carcinogenicity ddta.; Biometrics. 33:437.
                       Doll, R. 1971. Weibull distribution of
                      cancer Implications for models of
                      carcinogenesis. Jour. Roy. Slat. Sec. A. 13:
                      133.
                       Dripps, R.D., et al. 1977. Introduction  to
                      Anesthesia, the Principles of Safe Practice.
                      5th ed. WJ). Saunders Co., Philadelphia,
                      Pennsylvania, p. 121.
                       Druckrey, H. 1967. Quantitative Aspects of
                      Chemical Carcinogenesis, ln:R. Turhaut (ed.).
                      Potential Carcinogenic Hazards from Drugs,
                      Evaluation of Risks. UICC Monograph Series,
                      Vol. 7. Springer Verlag. New York.
                       Federation of American Societies for
                      Experimental Biology. 1974. Biological Data
                      Books. 2nd ed. Vol. III. P.L Aitman and D.S.
                      Dittment (eds.) Federation of American
                      Societies for Experimental Biology. Bethesda,
                      Maryland. Library of Congress No. 72-67738.
  Guess, H.. et al. 1977. Uncertanlty
estimates for low dose rate extrapolations of
animal carcinogenicity data. Can. Res. 37:
2475.
  Interagency Regulatory Liaison Gfoup.
1979. Scientific bases for identifying potential
carcinogens and estimating their risks.
February 6.
  International  Commission on Radiological
Protection. 1977. Recommendation of the
ICRP. Publ. No.  28. adopted January 17,1077.
Pergamon Press. Oxford. England.
  Mantel. N. and M.A. Schneiderman. 1975.
Estimating "safe lavels." a hazardous
undertaking. Can. Res. 35:1379.
  McNamara, B.P. 1976. Concepts in Health
Evaluation of Commercial and Industrial
Chemicals. In: MA. Mehlman, et al. (eds.),
Advances In Modern Toxicology, Vol. 1, Part
1. Wiley and Sons, New York. p. 61
  National Academy of Sciences. 1977.
Drinking Water and Health. National
Academy of Sciences, Washington. D.C.
  National Cancer Institute. 1976. Guidelines
for carcinogen bioassay in small rodents. NCI
Carcinogenesis  Technical Report Series, No.
1. February, 1978.
  Peto, R. 1973.  Weibull distribution for
continuous-carcinogenesfs experiments.
Biometrics. 29: 457.
  Stephen. C.E. 1980. Memorandum to J.
Slara. U.S. EPA. July 3.
  Stokinger, M.E. and RX. Woodward. 1958.
Toxicologic methods for establishing drinking
water standards. Jour. Am.  Water Works
Assoc. 50: 517.
  U.S. EPA. 1978. National Interim Primary
Drinking Water Regulations. Office of Water
Supply. EPA 570/9-73-003. U.S.  Environ. Prot
Agency.
  U.S. EPA. 1980. Seafood consumption data
analysis. Stanford Research Institute
International, Menlo Park, California. Final
report. Task 11. Contract No. 68-01-3887.
  Velth. G.D. 1980. Memorandum to C.E.
Stephen. U.S. EPA April 14.
  Veith. G.D., et al. 1979. Measuring and
estimating the bioconcentration factors of
chemicals in fish. Jour. Fish Res. Board .Can.
38:1040.                            •

Health Effects Methodology—Authors,
Contributors and Reviewers
Jerry F. Stara, ECAO-Cln, U.S. Environmental
  Protection Agency
Elizabeth L. Anderson. OHEA. U.S.
  Environmental Protection Agency
Roy E. Albert N.Y. University Medical
  School. New York. NY
Patrick R. Durkln, Syracuse Research
  Corporation, Syracuse, NY
Steven D. Lutkenhoff, ECAO, ITS.
  Environmental Protection Agency
Robert E. McGaughy, CAG, U.S.
  Environmental Protection Agency
Charles Ris, OHEA,  U.S. Environmental
  Protection Agency
Lawrence Anderson, CAG,  U.S.
  Environmental Protection Agency
Dolph Arnicar, CAG, U.S. Environmental
  Protection Agency
Steven Bayard.  CAG. U.S. Environmental
  Protection Agency
Edward Caiabrese, University of
  Massachusetts, Amherst. MA
Thomas Clarkson, University of Rochester,
  Rochester. NY  •

-------
                   Federal Register  /  Vol. 4S. No.  231 / Friday. November 28. 1980  / Notices	79357
Joseph Cotruvo. ODW, U.S. Environmental
  Protection Agency
Christopher DeRosa, ECAO-Cin, U.S
  Environmenlal Protection Agency
Kurt Ensleln. Cenesee Computer Center. Inc..
  Syracuse, NY
Jeffrey Gaba. OCC U.S. Environmental
  Protection 'Agency
Frank Gostomski. OWRS. U.S. Environmental
  Protection Agency
Daniel Greathouse, HERL-Cin, U.S.
  Environmental Protection Agency
Joseph Areas, Tulane University, New
  Orleans, LA
Dev Barnes, OWRS. U.S. Environmental
  Protection Agency
David L Bayliss, CAG, UJ. Environmental
  Protection Agency
Chao W. Chen, CAG, U.S.-Environmental
  Protection Agency
Herbert Cornish, University of Michigan, Ann
  Arbor, MI
Kenneth Crump. Science Research Systems,
  Int. Huston. LA
Michael  Doureon. ECAO-Cln. U.S.
  Environmental Protection Agency
John R. Pbwle HI. CAG, U.S. Environmental
  Protection Agency
R. John Gamer, HERL-Cin, U.S.
  Environmental Protection Agency
Terrence Crady, ECAO-Cin, U.S.
  Environmental Protection Agency
Bernard Habeiman, CAG, U.S. Environmental
  Protection Agency
Paul Hammond, University of Cincinnati.
  Cincinnati, OH
Charallngayya Hiremath, CAG, U.S.
  Environmental Protection Agency
Dinko KeJlo, Institute for Medical Research
  and Occupational Health, Zagreb, Yugo.
Chang S, Lao, CAG, U.S. Environmental
  Protection Agency
Marvin Legator, University of Texas Medical
  Branch, Galveston. TX
Jack McGinnlty, OTS. U.S. Environmental
  Protection Agency
Debdas Mukerjee. ECAO-Cin, U.S.
  Environmental Protection Agency
Ellen OTloherty, University of Cincinnati,
  Cincinnati, OH
Fred Passman, Energy Resources Co., Inc.
  Cambridge, MA
John F. Risher, ECAO-Cin. U.S.
  Environmental Protection Agency
Jeffery Rosenblatt, OTS, U.S, Environmental
  Protection Agency
Samuel Shibko, US FDA. Washington, DC
Harry T. Skalaky, Reynolds Metal Company
Todd W. Thorshuid, CAG. U.S.
  Environmental Protection Agency
Benjamin VanDuuren, N.Y. University
  Medical School, New York, NY
fames R. Withey, National Health and
  Welfare Canada Ottawa, Canada
Rolf Hartung, University of Michigan. Ann
  Arbor, Ml
Rudolf Jaeger, Harvard School of Public
  Health, Cambridge, MA
Arnold M. Kuzmack, CAG. U.S.
  Environmental Protection Agency
Si Duk Lee. ECAO-Cin. U.S. Environmental
  Protection Agency
Leland McCabe, HERL-Cin, U.S.
  Environmental Protection Agency
Myron Mehlman, Mobil OU Corporation, New
  York. NY
Jean Munson. University of Cincinnati.
  Cincinnati, OH
Nancy Othmer. OGC, U.S. Environmental
  Protection Agency
David I. Reisman, ECAO-Cin. U.S.
  Environmental Protection Agency
Paula K. Robereon, U.S. Environmental
  Protection Agency
H. Daniel Roth, Roth Associates
Dharm V. Singh, CAG, U.S. Environmental
  Protection Agency
Robert G. Tardiff. National Academy of
  Sciences, Washington, DC
Anne W.' TrontelL Energy Resources Co., Inc.,
  Cambridge, MA
John Van Ryzin, Columbia University, New
  York. NY
Ronald E Wyzga, Electric Power Research
  Institute

Appendix D—Response to Comments on
Guidelines For Deriving Water Quality
Criteria for the Protection of Aquatic
Life and Its Uses

Introduction
  Two versions of the Guidelines were .
published in the Federal Register for
comment. The first version (43 FR 21506,
May 18,1978 and 43 FR 29028. July 6,
1979) was simply published for
comment. The second (44 FR 15926.
March 15,1979] was published as part of
the request for comments on the water
quality criteria for 27 of the 65
pollutants. The second version was
meant to be clearer and more detailed
than the first, but very similar
technically. Since the two versions were
so similar, comments on both will be
dealt with simultaneously.
  Many comments were received that
no draft water quality criteria for any of
the 65 pollutants should have been
issued for public comment until the
comments on the first version of the
Guidelines had been dealt with
adequately and the Guidelines changed
appropriately. The comments on the first
version were read and the Guidelines
were revised in an attempt to make the
second version clearer and more
detailed than the first However, an
extensive revision of the technical
content of the Guidelines was not
attempted between the first and second
versions-because the Agency was
preparing water quality criteria based
on the Guidelines. The Agency could
have avoided this criticism simply by
not publishing any version of the
Guidelines for comment until March 15,
1978, but this would have greatly
reduced the length of time available for
people to consider the Guidelines and
comment on them. As it was, some
people commented that the comment
period announced on March 15,1979,
was too short.
  1. Comment—The procedures used  to
derive criteria in the "Red Book" were
 upheld in court and probably should still
 be used.
   Response—The procedures used in
 the Guidelines are similar to some of .the f
 procedures used to develop criteria in
 the "Green Book", "Blue Book", and
 "Red Book". The Guidelines are
 designed to be more objective and
 systematic, to deal more adequately
 with residues, and to incorporate the
 concept of a minimum data base.
   2. Comment—Criteria should be
 compilations of critically reviewed data
 with no synthesis or interpretation.
   Response—Neither P.L 92-500 nor the
 Consent Decree specify the form which
 a criterion must take. The Consent
 Decree (para. 11, p, 14) specifies that
 such criteria "shall state, inter alia,
 recommended maximum permissible
 concentrations". Adequate precedents
 have been set In the "Green Book",
 "Blue Book", and "Red Book" for the
 form of criteria used In the Guidelines.
   3. Comment—The Guidelines and
 criteria should be developed by a
 consensus of aquatic lexicologists rather
 than by EPA personnel only.
   Response—EPA certainly wants the
 Guidelines and the criteria to be as good
 as possible and as acceptable to as
 many interested people as possible. To
 this end, EPA has widely distributed
 draft versions of the Guidelines and the
 criteria documents, discussed them with
 many people, considered the comments
 received, and made many significant
 technical changes and editorial
 revisions. It is questionable whether or
 not a true consensus could have been
 reached by any means within the time
 available. In addition, EPA has a
 legislative responsibility which it should
 not delegate to someone else.
   4, Comment—The Guidelines should
 be updated regularly.
   Response—The Guidelines are not
 being promulgated as a regulation or
 directive. The purpose of presenting
 these Guidelines is to show how the
 water quality criteria for aquatic life
 were derived for the 65 pollutants. If
 EPA uses these  Guidelines again, they
 will be revised to take into account new
 data, concepts, and ideas.
   5. Comment—The objectives, purpose,
 and limitations of the Guidelines should
 be stated.
   Response—The introductory portion
 of the Guidelines has been expanded to
 address these subjects more fully.
   8. Comment—The Guidelines are too
 ambiguous.
   Response—The Guidelines have  been
 revised and rewritten, partly to improve
 clarity and provide additional details. It
 is not possible to provide explicit details
. on all items; in some areas only general
 guidance can be provided at this time.

-------
79356
                  Federal Register / Vol. 45. No. 231  / Friday. November 28. 1S80  /  Notices
EPA attempted to clearly and concisely
deal with all Issues which might
significantly affect the resulting criteria
without going into extreme detail on
every potential problem. Because
numerous judgments must be made, a
reasonable amount of experience in
aquatic toxicology will be necessary for
a person to utilize the Guidelines
effectively.
  7. Comment—The Guidelines are too
complex.
  Response—Deriving a water quality
criterion is a complex exercise because
several different kinds of data and a
wide variety of. organisms need to be
considered. In addition, because data
have been generated using various
procedures, numerous individual
decisions need to be made and the
Guidelines attempt to provide guidance
concerning decisions that seem to need
to be made frequently. The Guidelines
are more complex than initially
envisioned to help insure that criteria
for different pollutants are  derived in a
reasonably comparable manner.
Although the process of deriving a water
quality criterion for aquatic life is
complex, the Guidelines help organize
the process into logical components and
steps.
  8. Comment—The Guidelines should
be more flexible.
  Response—The Guidelines*are meant
to provide guidance  and at the same
time allow reasonable flexibility. They
have been used with quite a variety of
pollutants for which the requirements of
the minimum data base are satisfied,
and they seem to be reasonably
appropriate in all cases because the
experiences with these substances were
a major part of the basis for the
Guidelines. If sound scientific evidence
indicates that a particular aspect of the
Guidelines is not appropriate for a
specific substance, then some other
more appropriate procedure should be
used. However, the Guidelines should  •
not be changed based on individual
whim or personal preference.
  i. Comment—The Guidelines should
take into account synergism and
antagonism by a wide variety of factors
and the effect of the pollutant on
important ecological relationships.
  Response—Very little practically
useful information is available on these
factors in connection with the effects of
pollutants on aquatic organisms.
Synergism and antagonism are possible
between numerous combination of two
or more pollutants, and some data
indicate that such interactions are not
only species specific, but also vary with
the  ratios and absolute concentrations
of the pollutants and the life stage of the
species. Pollutants may affect the
 structure and function of aquatic
 ecosystems separate from their effects
 on individual species,  but practical
 applications of such ideas seem very
 tenuous at this time. Little information is
 available concerning such effects, and
 the significance of the available data is
 questionable. An obviously important
 ecological relationship is the
 dependence of higher organisms on
 lower organisns for food. Even here, the
 existence of numerous lower species
 and their adaptability reduces the
 Importance of any individual food
 species.
   10. Comment—The Guidelines should
 take into account all identifiable
 effects—beneficial as well as harmful.
   Response—Few tests have been
 conducted to identify beneficial effects
 of individual pollutants on aquatic
 organisms. However, beneficial effects
 are sometimes observed In chronic
 toxicity tests at concentrations below
 those that cause  adverse effects. Usually
 in such cases the organisms In low
 concentrations of the pollutant are
 longer or heavier or reproduce more that
 do the controls. Even if such effects are
 statistically significant, they are not
 Judged as adverse or harmful. On the
 other hand, a beneficial effect on one
 species may ultimately be to the
 detriment of a community if a balance
 between species is disturbed. Also, a
 concentration that benefits one species
 may harm a more sensitive species,
   11. Comment—The Guidelines should
 take into account analytical
 methodology.
   Response—The Guidelines do take
 into account analytical methodology in
 the definition of the substance, when
 necessary, but not In deriving the
.numerical value of the criterion.
 Concentrations which cannot be
 routinely measured accurately can often
 be measured accurately by nonroutine
 methods and, more importantly, do
 sometimes adversely affect aquatic
 organisms, When aquatic organisms are
 more sensitive than routine analytical
 methods, the proper solution Is to
 develop better analytical methods, not
 to underprotect aquatic life. One use of
 criteria should be to identify needs in
 analytical chemistry.
   12. Comment—The Guidelines should
 take into account (a) production and
 usage patterns, (b) chemical, physical
 and biological factors pertaining to
 degradation and fate of pollutants,
 Including properties such as solubility in
 water, decay rate, persistence, and
 transformation pathways, and (c)
 whether or not a criterion is needed for
 the substance.
   Response—Items included in (a} end
 (b] may be important in deciding
whether a criterion is needed for a
substance, but the Guidelines are
intended to be used after the decision
has been made that a criterion is *
needed. EPA is presently developing
principles that can be used to decide
whether or not a criterion is needed for
a substance and items such as those
listed above are  probably some of the
factors that should be considered when
deciding whether or not a criterion is
needed. If the toxicity of the chemical is
used to evaluate the need for a criterion,
the Guidelines may be useful in the
collection and interpretation of the
available toxicity data.
  13. Comment—The Guidelines should
take into account costs to states and
industries, technological feasibility, and
such characteristics of bodies of water
as assimilative capacity, dispersal,
dissipative factors, dilution, hydrology,
mixing zones, and sediment.
  Response—Factors such as these
should be considered in developing
standards, but not in deriving criteria.
EPA is presently developing an
implementation policy which will
describe which of the above factors and
which characteristics of the pollutant
should be used, and how they should be
used, in developing standards.
  14. Comment—The Guidelines are not
appropriate for establishing a
concentration which may be present in
an effluent.
  Response—The Guidelines are for
deriving water quality criteria, not
effluent standards nor mixing zone
standards nor water quality standards.
Water quality criteria will probably be
one factor taken into account in the
development of water quality standards
and toxicity-based effluent standards,
but not technology-based effluent
standards. EPA is presently developing
policies concerning proper use of water
quality criteria in various regulatory
activities.
  15. Comment—The derivation of
criteria should be fundamentally a
scientific exercise and should not
employ subjective judgments.
  Response—No exercise which
involves the use and interpretation of
data can avoid subjective Judgment.
Indeed, even the generation of scientific
data requires subjective Judgment, such
as how many test organisms to use,
what temperature to use, etc. One may
decide to accept the recommendations
of experts, but this is usually still a
subjective decision. In statistics the
subjective decisions are made on the
basis of probability statements but the
final decisions are still subjective
judgments. Although the development of
the Guidelines and the derivation of
criteria cannot avoid subjective

-------
                   Federal Register / Vol. 45. No.  231 / Friday, November 28. 1980 / Notices
                                                                       79359
 decisions, gross extrapolations, wild
 assumptions, and novel judgments can
 be avoided. One can also avoid using
 large safety factors to "make up" for
 insufficient data. When some agreement
 exists between experts, such as on test
 temperature and duration of tests, the
 collective opinion can usually be used.
 EPA feels that the Guidelines do not go
 loo far beyond the state-of-the-art and
 do not produce criteria by extrapolating
 far beyond the usefulness of the data.
   16. Comment—The Guidelines should
 not use unproven extrapolations.
   Response—EPA feels that the
 extrapolations used in the Guidelines
 are reasonable for most pollutants.
 Probably the most questionable
 extrapolation is the acute-chronic ratio.
 but even here an arbitrary ratio is not
 used. Indeed, the ratio used is usually a
 mean of experimentally determined  .
 acute-chronic ratios for at least three.
 not |ust one, species. In addition,  the
 species must include at least one  fish
 and one invertebrate. Even this amount
 of data does not "prove" the validity of
 the extrapolation, but it should provide
 reasonable evidence for or against the
 use of the ratio with any particular
 substance. To achieve reasonable
criteria without using any extrapolations
would require acute and chronic tests
with many more species. This would be
a high price to pay for disallowing any
use of scientific inference in deriving
criteria.
  The early versions of the Guidelines
used adjustment factors and sensitivity
factors which were averages derived
•from data for a wide variety of
substances and thus were attempts to
make some extrapolations across all
substances. The present version of the
Guidelines is based on a minimum data
base for each individual pollutant and
 the calculations are essentially
pollutant-specific. Thus no
extrapolations are made from one
pollutant to another.
  17. Comment—Laboratory tests
 3verestimate the toxicity of materials
 Decause the test organisms are  stressed
 ay the artifical conditions.
  Response—Laboratory conditions
 :ertainly are artificial, but they do not
 lecessarily stress the test organisms.
 Drganisms which survive, grow, and
 eproduce well in the laboratory cannot
 >e stressed too much. Organisms in a
 aboratory might be considered
 ampered because they do not have to
 ompete for food and are not subject to
 tress due to predators and changing
 nd extreme conditions of turbidity,
 ->mperature, flow, and water quality.
 .Iso. laboratory organisms are rarely
 jbject to stress from pollutants. Some
 oecies probably have longer average
 life spans in-laboratories than they do in
 field situations.
   18. Comment—Laboratory tests
 underestimate the toxicity of materials
 because the tests are usually conducted
 with species which are hardy,
 adaptable, and insensitive.
   Response—Species which are readily
 adaptable to laboratory conditions are
 not necessarily insensitive as evidenced
 by the great range of sensitivities
 obtained in laboratory tests for some
 individual pollutants with different
 species. In fact, once the the proper .
 techniques are developed, a wide
 variety of species can survive, grow, and
 reproduce well in laboratories. When
 the proper techniques are discovered
 and a species changes form "difficult" to
 "easy", its sensitivity does not change.
 Also, some species and life stages which
 are fragile and must be handled with
 great care are  not particularly sensitive.
 On the other hand, because so few
 species have actually been tested  in
 laboratories, species which are more
 sensitive than any of those tested  in
 laboratories, species which are more
 sensitive than any of those tested
 probably exist for most substances.
  19. Comment—Laboratory tests are
 artificial and contrived and do not
 represent the real world.
  Response—Laboratory tests are
 indeed artificial but they are not
 contrived to give results that are
 unnecessarily high or low. Organisms in
 a laboratory are generally acclimated lo
 water and conditions of constant and
 desirable quality, whereas in the field
 (hey are often subjected to fluctuations
 and extremes. Organisms in a
 laboratory do not have to compete for
 food and are not subject to predators or
 pollution. Organisms In the field are
 often exposed to more than one
 pollutant at  a time, with the
 combinations and concentrations
 changing often.
  It is true that aquatic organisms are
 usually exposed to instantaneous high
 concentrations in laboratory tests, but in
 field situations organisms are often not
 given much chance to acclimate to spills
 or short-term discharges. Also, some
 ameliorating effects occur in field, but
not laboratory, situations, but such
effects are not always dependable over
 long periods of time. The concentrations
 of mitigating anions, suspended solids,
 and complexing agents are relatively
 constant in some bodies of water, but
not in others. Suspended solids probably
 do sorb and detoxify significant
 amounts of some pollutants,  but high
concentrations of suspended solids also
 stress some  aquatic organisms. In
addition, organisms are usually fed in
chronic tests, so the test solution
contains suspended solids and dissolved
organic carbon from the food and fecal
matter. Degradation and other
transformations are more likely in field-
situations than in laboratory situations.
but degradation products are not always
less toxic than the undegraded material.
On the other hand, many of these kinds
of considerations will probably be taken
into account when site-specific criteria
and standards are developed under the
implementation policy which is being
developed by EPA.
  20. Comment—Laboratory tests are
poor predictors of what will happen in
field situations.
  Response—If conditions are
comparable, laboratory toxicity tests are
useful predictors of what will happen in
field situations. The usefulness of such
predictions will depend on how
carefully one accounts for differences
between species, water quality, and the
form of the pollutant. Extrapolations are
much more difficult for some pollutants
than for others. Water quality affects the
toxicity of some pollutants much more
than others, and species differences,
even within families, are much greater
for some pollutants than for others. If
such factors are taken into account,
useful-predictions are possible. In what
is probably the most extensive
comparison available of laboratory and
field data (Geckler, J. R., et aL 1978.
Validity of Laboratory Tests for
Predicting Copper Toxicity in Streams.
EPA-600/3-7&-118. U.S. EPA. Duluth,
MN 208 pp.), It was found that  effects
observed in laboratory exposures were
also observed in field exposures.
However, avoidance, which was not
studied in laboratory exposures, was
observed in the field exposures.
Laboratory to field  comparisons are not
simple because several factors must be
taken into account, the laboratory test
must be conducted well and the field
observations and measurements must be
extensive. Although adverse effects
observed in laboratory tests will usually
occur in similar field situations, a
problem exists with the bioaccumulation
of some persistent substances. For
example, PCB's seem to bioaccumulate
to much higher levels in some bodies of
water than they do in laboratory tests.
  21. Comment—The Guidelines should
place more emphasis on field
information than on laboratory
information.
  Response—Field information on
effects of pollutants on natural
populations is acceptable, but the  -
collection of definitive information  of
this type is high risk and costly. Few
studies on the effects of pollution on
natural populations provide definitive
information because of the multitude of

-------
79360
Federal  Register / Vol.,45. No. 231  / Friday, November 28. 1980 / Notices
variables that need to be taken Into
account. The major advantage of Held
studies la that conditions are natural
(I.e., conditions are not controlled), but
this la also the major problem with field
studies. With uncontrolled conditions,
numerous variables must be taken Into
account, because any individual
variable or combination of variables
may affect the results or Indeed may be
the cause of. the results. Therefore, field
studies on natural populations usually
must last over several seasons and
possibly over more than one year to be
reasonably sura that proposed cause-
and-effect relationships are real.
  Another problem with field studies
that are based on statistically significant
differences is the power of the test
Because natural biological, spaciaL and
temporal variability is often rather great,
a large number of samples is usually
required to detect even a moderate
change. A field study, which purports to
show that no change occurred is of no
value if the power of the test calculated
from the experimental design and
observed variability was not high
enough.
.  Because field studies are Ugh cost-
high risk ventures, well-designed
laboratory testa are usually much more
cost-effective for obtaining data on (1)
the toxlcity of substances to a variety of
species and (2) the effect of various
water quality characteristics on toxlcity.
Laboratory testa have been shown to
generally be useful predictors of what
happens in a field situation, and so it
makes little sense to conduct high risk,
high coat field studies rather than
laboratory tests. Even definitive field
studies rarely provide enough
information to allow extrapolation of
results to other situations, so field
studies are more useful in reviewing
criteria than in deriving criteria.
  22. Comment—Field verification of
laboratory tests and of the Guidelines
are needed.                         •
  Response—Field verification of
laboratory tests and of the Guidelines
are certainly desirable and provide
Information that cannot be obtained In a
laboratory. Field verification studies do
not need to be as risky or as costly as
studies on the effects of a pollutant on
natural populations because verification
studies can be designed (I) as a slde-by-
side comparison of the results of
laboratory tests and field tests or (2)
based on existing results of laboratory
tests.
  23. Comment—EPA should allow
criteria to be derived using on-site acute
toxicity tests and an application factor.
  Response—This approach is usually
suggested for developing effluent
standards but may be just as applicable
                     to deriving water quality criteria under
                     certain conditions. This approach
                     cannot be used with pollutants whose
                     most sensitive adverse effect is due to
                     residues. Also, it can only be used when
                     the application factor has already been
                     acceptably determined. Finally, acute
                     tests must be determined with either an
                     appropriate range of species or with an
                     appropriate sensitive species. The
                     implementation policy presently being
                     developed by EPA will probably allow
                     the use of appropriate on-site toxicity
                     tests In the envelopment of site-specific
                     criteria and standards.
                       24. Comment—It Is not dear what
                     level of protection is intended.
                       Response—EPA feels that it is not
                     possible to specify a minimum level of
                     protection that is necessary to "protect
                     aquatic life" or even to protect a
                     particular species for such reasons as:
                       a. There are so many untested
                     species.
                       b. Little practically useful information
                     is available concerning synergism,
                     antagonism, ecological relationships,
                     and avoidance.
                       c. The effect of factors such as
                     temperature on toxlcity seems to be
                     species-specific for at least some
                     substances.
                       d. Information is not available
                     concerning what amount of any effect
                     would be ecologically significant and
                     whether the amount is species-specific.
                       One possible conclusion is that to
                     protect aquatic life, ail species must be
                     adequately protected. A possible
                     extension of this would be that all
                     criteria should be zero because any
                     amount of any pollutant may affect
                     some aquatic organism. Indeed, the
                     assimilative capacity of body of water
                     largely depends on the  ability of aquatic
                     life to "process" pollutants and to some
                     extent, any organism which "processes"
                     a pollutant is in some way affected by it
                       The apparent level of protection is
                     different for each kind of effect (acute
                     toxlcity to animals, chronic toxicity to
                     animals, toxicity to plants, and
                     bioaccumulation) because of the quality
                     and quantity of the available
                     information. An attempt was made to
                     take into account such things as the
                     importance of the effect, the quality of
                     the available data, and the probable
                     ecological relevance of the test methods.
                     Thus it was felt that with regards to
                     toxicity to animals it was probably not
                     necessary to protect all of the species all
                     of the time, but it certainly seems
                     appropriate to protect most of the
                     species most of the time and to protect
                     important species.
                       On the other hand, the data base on
                     toxicity to aquatic plants is usually very
                     small and a variety of tests and
endpoints have been used, especially
with algae. Also, little information is
available concerning the ecological
relevance of the results of any toxicity
test with algae in a concentrated test
medium, especially because so many
species of algae exist in each body of
water.
  The results of bioconcentration tests
with organic chemicals, but not with
inorganic chemicals, can apparently be
extrapolated reasonably well based on
percent lipids from one aquatic animal
species to another, at least within
commercially and recreationally
important species. In addition, die limits
on acceptable concentrations In tissue
are reasonably well defined in some
cases,
  These kinds of considerations merely
illustrate the complexity of the problem
and the necessity for making decisions
about each kind of effect individually. In
addition, it is important to distinguish
between the apparent level of protection
provided by the Guidelines and the
actual level of protection which will
result in a field situation from the use of
the implementation policy.
  No attempt was made to develop
Guidelines which would achieve a
predetermined numerical level of
protection. For each effect much
desirable information is not available,
and so it would be misleading to imply a
level of sophistication that is not
currently possible. EPA believes that the
present state-of-the-art In aquatic
toxicology does allow some useful
conclusions about the ability of a
substance to adversely affect aquatic
organisms and their uses whenever the
requirements of the minimum data base
are satisfied, with the full realization
that the resulting criterion may be
somewhat overprotective or
underprotective.
  In almost all cases more data would
be desirable and so an attempt to reach
the "golden mean" will sometimes result
incriteria being to high and sometimes
too low. One alternative is to derive no
criteria until all desirable data are
available; this is unacceptable because
it will almost always result in no criteria
and no protection. The other alternative
is to apply safety or uncertainty factors
that are inversely proportional to the
adequacy of the data base. In the long
run this approach would encourage the
generation of useful data where it was
most  needed, but in the short run would
require many significant subjective
decisions beyond the Current state-of-
the-art
  25.  Comment—The Guidelines should
not base criteria on '-worst case"
assumptions.

-------
                   Federal  Register / Vol. 45. No.  231 / Friday. November  28, 1980  /  Notices             79361
   Response—The phrase "worst case
 assumptions" usually refers to the
 assumption that both the worst water
 quality and the most sensitive life stage
 occur at all times. These two
 assumptions are a natural result of the
 two concepts that criteria should be
 constant throughout the year and that
 aquatic life is not adequatley protected
 if it is not adequately protected
 throughout the year. The implementation
 policy being developed by EPA will
 determine whether site-specific criteria
 must be constant throughout the year. If
 not then the "worst case assumptions"
 will not apply. Although the Guidelines
 might be viewed as making the "worst
 case assumptions", the implementation
 policy will determine whether the site-
 specific water quality criteria and
 standards will be based on these
 assumptions,
   26. Comment—Safety factors should
 be used to protect against such things as
 potential subtle, but important, long
 term effects.
   Response—Pollutants may cause
 many direct and indirect adverse effects
 which have not been studied
 adequately. For instance, some
 substances may make aquatic organisms
 more susceptible to disease or other
 stresses. In spite of such possibilities,
 the available information indicates  that
 the major possible adverse effects are
 covered in the Guidelines and that
 adequate protection will usually be
 achieved without the use of safety
 factors. Safety factors would certainly
 offer additional protection, but the
 available information: does not show  *
 that significant additional protection is
 needed.
  Safety factors of from 10 to 1000 are
 often used  to protect people mainly
 because people feel that people are
 more important than aquatic organisms
 and because humans are usually
 protected on the basis of tests with
 other species of animals/ thus resulting
 in a greater uncertainty in the
 applicability of the results. Complete
 protection can only be achieved by
 setting all criteria at zero. Unfortunately.
 even "Mother Nature" sometimes
 seriously harms large groups of aquatic
 organisms, such as during droughts  or
.severe winter freezes. EPA feels that   .
 complete protection is neither feasible,
 desirable, nor possible. In addition,
 aquatic ecosystems can recover from
 some'adverse effects.
  27. Comment—The Guidelines do not
 provide for an adequate margin of
 safety.
  Response—If "margin of safety" is
 interpreted to mean  "safety factor", then
 ihe Guidelines do not provide  a margin
 of safety. If the Guidelines are viewed
 as deriving criteria for a constant
 quality water, then they provide a
 margin of safety during those portions of
 the year during which the most sensitive
 life stage does not occur. Although some
 species may occasionally be adversely
 affected. EPA feels that the Guidelines
 provide adequate safety because
 aquatic communities and their uses
 should not incur any substantial or
 permanent damage. Whether or not site-
 specific criteria will have a margin of
 safety will depend on how they are
 derived;
   28. Comment—Criteria should be set
 at the least restrictive concentration and
 states can then apply more restrictive
 concentrations when necessary.
   Response—It is unclear what is meant
 by the "least  restrictive concentration"
 but presumably it would be a
 concentration which would not protect
 very many aquatic communities and •
 their uses. This is contradictory to the
 concept that criteria are to protect
 aquatic life and its uses; The
 implementation policy being developed
 by EPA will allow site-specific criteria
 to be higher or lower than the criteria
 derived using the Guidelines, when
 adequate Information is available.
  29, Comment—The Guidelines should
produce criteria in the form of a
concentration-risk curve with
appropriate confidence limits for each
kind of effect.
  Response—EPA feels that a risk
analysis approach is certainly desirable,
but far beyond the state-of-the-art at
this time. When dealing with safety to
humans, only one species is being
 protected and extrapolations are made
far outside the limits of the actual test
results, such as to 1 death in 100,000
people. With aquatic life, numerous
species need to be protected and
extrapolation far beyond the actual data
is not readily  accepted. In addition,
safety or uncertainty factors are more
readily accepted when protecting people
than when protecting aquatic organisms.
  Most aquatic lexicologists are not
willing to.let criteria for the protection
of aquatic life be as dependent on,
mathematical models, assumptions, and
manipulations as on the actual test
results. Most people with experience in
aquatic toxicology have an intuitive
"feel" about how data should be
interpreted and the Guidelines are
merely an attempt to formalize a
resaonable approach. The Guidelines
could be written as mathematical
algorithms and some approach such as
error models could be developed in
order to derive confidence limits.
However, the algorithms and models
would contain many unproven
assumptions and, to be worthwhile.
would undoubtedly require more data
than are usually available. Although
such models and algorithms would be
acceptable to many statisticians and
may be an appropriate future goal, the
current Guidelines need to be uscoble
by and comprehensible to current
aquatic lexicologists. Most experienced
aquatic lexicologists will judge the
reasonableness of any set of Guidelines
by comparing the resulting criteria for
various pollutants with the data
available for those pollutants using a
"common sense" interpretation of data.
  30. Comment—The Guidelines should
not use unsound statistical procedures
or misuse sound statistical procedures.
  Response—EPA has tried to make
sure that no statistical procedures are
misused in the Guidelines, that no
unsound stalislical procedures arc used,
and that the purposes of the calculations
are explained adequately.
  31. Comment—It appears that
geometric means were used instead of
arithmetic means in the Guidelines to
obtain lower values.
  Response—Decisions such as this
were made throughout the Guidelines on
a case-by-case basis, and none were
based on whether the resulting criteria
would be higher or lower. The selection
of the procedure used to calculate the
mean could be based on the distribution
of the values in the individual data set.
Unfortunately, with small data sets
rarely is it possible to reject many
possible distributions and with large
data sets all possible distributions are
often rejected. Because many of the data
sets of interest in the Guidelines are
small, a reasonable approach is to base
the selection of a procedure for
calculating the mean on some general
principles such as:
  a. Sets of ratios and quotients are
likely to be closer to lognormal than
normal distributions. Thus geometric
means, rather than arithmetic means,
are used for acute-chronic ratios and for
bioconcentration factors.  '.
  b. When there are numerous
independent possible sources of error
for each datum In a set, the error tends
to be multiplicative rather than additive.
Thus when the acute or chronic toxicity
of a substance to a particular species is
determined; in different laboratories
using different batches of organisms,
different waters,  etc, the geometric
means should be used to calculate the
species mean value rather than the
arithmetic mean.
  c. If a set of numbers approximates a
lognormal distribution, the logarithms of
the numbers will approximate a normal
distribution.
  d. The distribution of the sensitivities
of individual organisms in a toxicity test

-------
79362
                  Federal Register / Vol. 45, No. 231 / Friday. November 28. 1980 / Notices
is likely to be closer to a lognormal
distribution than a normal distribution.
Thus the geometric mean, rather than
the ari theme tic mean, of the upper and
lower chronic limits is used.
  32. Comment—There should not be
any, criteria which apply to all bodies of
water. Criteria should be specific for
Individual states, regions, other
geographic areas, or bodies of water.
  Response—The Guidelines are
designed to provide guidance In the
collection and Interpretation of data
concerning the effects of pollutants on
aquatic life and its uses. The uses of the
resulting criteria  will be described by
EPA in various regulations. If desired,
the Guidelines can be appropriately
modified and used to derive a criterion
specific to one or more bodies of water
or geographic areas if an appropriate
data base is available. The critical
literature reviews on which the criteria
are based will be available for use In the
derivation of local, state, or regional
criteria. The latitude allowed for
deriving local, state, or regional criteria
and standards will be determined by the
implementation policy presently being
developed by EPA.
  33. Comment—The Guidelines should
result in criteria that are specific for
individual species or groups of species
(e.g.. warmwater and coldwater).
  Response—If the necessary data were
available, criteria could be  derived for
any particular species or group of
species. It was Impractical for EPA to
derive criteria for many such groups, but
a relatively simple division is freshwater
and saltwater organisms because these
two groups rarely coexist Most other
possible general divisions of species are
faced with the problem that species
coexist in various combinations unless
the  groups are very harrow. In addition,
toxicity data are rarely available for -
very many individual species and so
data for representative species must be
used, unlasa appropriate new data are •
generated. Also, the available data
sometimes show  wide differences within
families so extrapolations from one
species to another are often tenuous.
Because of these  problems, deriving
criteria for individual species or groups
of species was deemed impractical.
  34. Comment—A criterion should be
one number, not two.
  Response—The two-number criterion
is an acknowledgement that aquatic
organisms  can tolerate short exposures
to concentrations that are higher than
those they can tolerate continuously. In
a two-number criterion, the higher
number can assure that  short-term
fluctuations above the average are not
too  high, whereas the lower number  can
assure that the long-term average is not
 too high. A one-number criterion could
 be derived by using the existing 24-hour
 average as an instantaneous maximum.
 This would certainly provide additional
 protection, but would provide
 unnecessary overprotection in most
 cases. Because a one-number criterion
 would be more of an approximation
 than a two-number criterion, one-
 number criteria would be too high or too
 low more often and to a greater degree
 than two-number criteria.
   35. Comment—The criteria should not
 specify sampling schemes.
   Response—Criteria should state
 numerical concentration limits in terms r
 of exposure durations because,
 everything else being constant the
 amount of adverse effect depends on
 both the concentration of the pollutant
' and the duration of exposure. Criteria in
 the Green Book, Blue Book, and Red
 Book were usually stated as single
 numbers with no duration expressly
 stated. The implication was that the
 criteria were never to be exceeded at
 any time. Each criterion was apparently
 and instantaneous maximum. In
 practice, however, standards derived
 from these criteria were usually
 enforced on the basis of 24-hour
 composite samples. To avoid any
 ambiguity, the Guidelines specify that a
 criterion should be explicitly stated in
 terms of two time frames: an
 instantaneous maximum and a 24-hour
 average. However, this is not a
 specification for a sampling scheme.
 Standards developed from such a
 criterion should probably specify a
 sampling scheme for compliance
 monitoring, but it would not necessarily
 be in terms of point measurements and
 24-hour averages.
   Any sampling scheme used to
 determine whether or not an ambient
 concentration exceeds a water quality
 criterion or a comparable water quality
• standard should take into account such
 things as the ratio of the instantaneous
 maximum and the 24-hour average and
 the retention time of the body of water
 because these will primarily determine
 which portion of the criterion is most
 limiting in any specific situation. The
 sampling scheme should probably also
 take into account the cost of the
 analyses and results of any past .
 analyses.
   36. Comment—The criteria should be
 stated in terms of time frames longer
 that an instantaneous maximum and a
 24-hour average.
   Response—These two time frames
 were chosen because they would allow
 the derivation of a criterion which
 would be less restrictive than, but just
 as protective as, the previous one-
 number criterion. These two specific
time frames were chosen because they
match two kinds of samples that are
commonly collected: grab samples and
24-hour composite samples. THese
specific time frames could probably be
changed somewhat without much
practical effect but EPA saw no
particular advantage to anyone to
introducing novel time periods. For
example, for all practical purposes in
most situations a 10-minutes average is
probably about the same as an
instantaneous maximum.
  Large increases in the time frames,
however, would not provide the same
amount of protection. If the
instantaneous maximum were changed
to a 24- or 98-hour average, and the 24-
hour average were changed to a 7- or 30-
day average with no change in the
numerical limits, the amount of
protection afforded aquatic life would
fall to an unacceptable level. The longer
the time span for the average, the higher
the instantaneous concentration could
be for short periods of time within that
span. Although most chronic tests last
for 28-days or longer, some chronic
effects may be caused by short
exposures of sensitive life stages. If the
acute-chronic ratio is small fluctuations
in the instantaneous concentration may
even cause acute toxidty, especially for
cumulative pollutants, because for some
substances the 24-, 48-. and 98-hour
acute values do not differ too much.
  37. Comment—A two-number
criterion will be difficult to enforce.
  Response—Criteria'are not
enforceable. Standards are enforceable.
When standards to protect aquatic life
are developed, they may or may not be
in the same format as the criteria for
aquatic life. Few standards are
adequately enforced because of the high
cost of continuous monitoring. The real
value of many criteria and standards  is
in the design of waste treatment
facilities; a two-number criterion should
be a better basis for design than a one-
number criterion.
  38. Comment—The criteria should be
expressed to one significant figure, not
two.
  Response—EPA acknowledges that
there is much variability in some of the
data and that the range of sensitivites is
often great When the requirements of
the minimum data base are satisfied and
the data agree reasonably well, two
significant figures are not unreasonable.
Rounding off to one significant figure
could arbitrarily raise or lower the
criterion by up to forty percent with no
apparent consistent benefits to
dischargers, regulators, or aquatic life.
  39. Comment—The Guidelines should
only use data for species that ought to
be protected.

-------
                   Federal Register  /  Vol. 45. No. 231  / Friday. November 28. 1980  /  Notices
                                                                      79363
  Response—In order to protect
 commercially and recreationally
 important species, a wide variety of
 "unimportant" species must also be
 protected. Such so-called "unimportant"
 species include the food organisms all
 the way to the bottom of the food chain.
 The "important" species in an aquatic
 community cannot maintain themselves
 without the help of primary producers,
 primary consumers, nitrlfiers,
 dentrifiers, detritivores and saprophytes.
  40. Comment—Criteria should not be
 based on sensitive, short-lived-
 invertebrates.     \
  Response—Many species of
 invertebrates are short-lived and are not
 widely distributed. However, these
 numerous short-lived, local species do
 serve important functions and should be
 represented In the data base. This group
 of organisms needs to be protected even
 if no one species can be considered
 important.
  41. Comment—Criteria should protect
 endangered species.   -
  Response—EPA agrees that criteria
 should protect endangered aquatic
 species. However, very few toxicity
 tests have been conducted with
 endangered species, and it does not
 appear feasible to require tests with
 such species. Endangered species are
 some of the many untested species
 which should be protected by criteria
 derived from available data using the
 Guidelines.
  42. Comment—Migratory species are a
 special problem.
  Response—Migratory species should
 usually be protected by criteria derived
 using the Guidelines unless such species
 are unusually sensitive. Migratory
 species may be especially susceptible to
avoidance, but few data are available to
compare species on this basis.
Avoidance may be a serious latent
problem because it might apply to all
motile species, rather than just
migratory species, and it has not been
studied very much.
  43. Comment—Estuarine species were
 ignored.
  Response—The term "saltwater
organisms" is meant to include estuarine
species as well as true marine species.
  44. Comment—The classification
"invertebrates" includes species that are
too  dissimilar to be grouped together.
These species should be separated into
phyla or classes.
  Response—The never-ending
arguments between the "lumpers"  and
the  "splitters" can only be resolved by
considering the advantages and
disadvantages of each approach in each-
situation. The "splitters" can usually
argue that  obvious differences should be
taken into  account and it is certainly
true that shrimp are different from
insects and both are different from
worms. It can also be argued that there
are significant differences within phyla,
classes, and families. Each species could
be considered a separate group, if
differences between stains are
arbitrarily ignored. After the species are
split into separate groups, the problem
then would be whether to recombine the
data to derive one criterion for all
species or to derive one criterion for
each group. If numerous criteria are
derived for a pollutant, how are these  to
be used to develop standards? Another
problem is that unless more data are
generated, the greater the number of
groups, the less information there is
available per group.
  The basic question is "What are the
important differences that need to be
taken into account and how should this
be done?" Because there are differences
between taxonomlc groups, the
Guidelines require data on a number of
species from a varitety of taxonomic
groups. The information of each
separate  species is treated individually.
This approach preserves the differences
between  species and allows all species
to be considered in the development of
the criterion. The number of data points
is increased and the range of the data  is
readily apparent Because
"invertebrates" is already a large
diverse group and because the range of
sensitivities of fish usually overlaps that
of invertebrates, little justification exists
for not combining all aquatic animals.
  45. Comment—Do not extrapolate
from freshwater organisms to saltwater
organisms or vice versa.
  Response—Criteria and absolute
toxicity values were not extrapolated
from fresh water to salt water, but some
relative data were, when it did not
appear that factors such as salinity
affected the  data. The toxicity of some
substances apparently is significantly
affected by salinity, but most substances
seem to have overlapping ranges of
toxicity to freshwater and saltwater
organisms. However, because these two
kinds of organisms rarely inhabit the
same body of water simultaneously,
separate criteria were derived for each.
Even though these two kinds of
organisms are physiologically different.
they do not seem to be too different   •
lexicologically. Bioconcentration factors
and acute-chronic ratios seem to be
fairly similar for many freshwater and
saltwater species for many pollutants.
particularly  organic chemicals.
  46. Comment—The Guidelines base
the criteria only on sensitive species and
do not take into account insensitive
species.
  Response—The Guidelines do not
necessarily base the criteria on the data
for the most sensitive species. However,
an aquatic ecosystem cannot be
protected by protecting only the species
which are insensitive. Protecting half the
species will probably not protect the
community. To offer reasonable
protection to aquatic life and its uses.
each major kind of organism and each
major use must be given reasonable
protection. In some cases it may in fact
be necessary to protect the most
sensitive species if it is a highly
desirable species.
  47. Comment—Species should be
tested at their environmental extremes.
  Response—Toxicity tests with each
pollutant could .indeeed be conducted
with some or all species under a variety
of extreme conditions and the lowest
result obtained with a species could be
used instead of a mean result On the
other hand, differences between results
with different species seem to be much
greater, and therefore more important,
than the differences between results
obtained with one species under
different conditions. Furthermore,
criteria need not necessarily protect
species from all stress under the most
extreme conditions, because aquatic
communities and populations  of
individual species can recover from
some perturbations.
  48. Comment—Only data for species
that are widely distributed.
representative, critical, indigenous,  '
important, ecologically relevant and
sensitive should be used.
  Response—Few species would satisfy
all of the requirements that have been
suggested. As more and more  data are
obtained with a wider variety of species
for any one pollutant, it becomes more
obvious that few if any species are
atypically sensitive, although that may
not be true for aquatic communities
which contain very-few species. No data
exist to show that species in any one
key role are lexicologically more
sensitive than other kinds of species.
Ecologically relevant species and
species that have key roles or are
relevant to the overall functioning of
viable ecosystems are not necessarily
lexicologically different from other
species. EPA feels that if the available
data cover an adequate number and
variety of species, it is not necessary to
try to identify and conduct tests with all
important, sensitive species. In addition,
the derivation of a criterion should not
be based only on  sensitive species,
because a knowledge of the range of
sensitivities may be useful. For instance.
elevated concentrations of a pollutant
that  produces a narow  range of species
sensitivities are likely to cause more

-------
79364
Federal  Register / Vol. 45, No. 231  / Friday. November  28. 1960 / Notices
damage than elevated concentrations of
a pollutant that produces a wide range
of species sensitivities.
  49. Comment—The distinction
between Icnizable and unionizable
compounds is not very good because
some chemicals ionize and reach
chemical equilibrium very slowly and
others very rapidly.
  Response—Most chemicals can .
readily be classified Into one of three
groups:
  A. Chemicals that ionize, including
hydrolyze,  at least 90% and reach 90% of
equilibrium in less than 8 hours in most
surface waters.
  B. Chemicals that ionize, including
hydrolyzfe, less than 10% in 30 days in
most surface waters,
  C. Chemicals that do not fit into either
one of the above categories.
  For the purpose of the Guidelines,
chemicals hi the A group should be
considered ionizable, chemicals in the B
group should be considered non-
ionizable, and chemicals in the C group
should be classified on a case-by-case
basis. Although the distinction between
ionlzabls and unionizable may not be
perfect, It is very useful for most
chemicals.
  SO. Comment—Each individual
organic compound should be considered
separately.
  Response—The vast majority of
organic chemicals will be considered
separately  according to the Guidelines
except for structurally similar organic
compounds that meet all three
specifications given in the Guidelines,
such as polychlorinated biphenyls and
toxaphene.
  51, Comment—In-stream water
quality criteria are meaningless for
substances lhat are highly insoluble.
  Response—The concentration of some
substances in sediment may be
important separate from the
concentration of the substance in the
ambient water and for these compounds
a sediment quality criterion may be
necessary. Generally such compounds .
can also cause adverse effects if the
concentration in the ambient water is
too high even if the concentration in the
sediment is low. Thus for such
compounds both kinds of criteria may
be necessary rather than just one or the
other.
  52. Comment—If a substance is not
dissolved, it is  not biologically or
lexicologically available.
  Response—Although this may usually
be true, it certainly does not apply to
elemental mercury which can be
oxidized and methylated to form a very
toxic compound. Some organic acids
and phenols and hydroxide and
carbonate salts of metals have
                     solubilities which differ substantially
                     from one body of water to another.
                       53. Comment—Criteria for metals
                     should not be for total metal.,
                       Response—Criteria for metals will
                     generally not be based on total metal.
                     Most will be based on total recoverable
                     motal because forms of metals that are
                     not measured in the total recoverable
                     procedure probably are not, and will not
                     become, toxic. A major problem is that
                     some people use a procedure for total
                     recorverable, but report the results as
                     total, metal. In many situations the two
                     results are about the same, but in some
                     cases the results are quite different, •
                       54, Comment—The Guidelines should
                     give .more guidance for distinguishing
                     between acceptable and unacceptable
                     data.
                       Response—The Guidelines contain as
                     much detail on this subject as EPA
                     believes is currently  feasible. Items such
                     as the maximum acceptable control
                     mortality and minimum number of test
                     organisms are based on what many
                     aquatic lexicologists generally feel are
                     acceptable, as expressed in published
                     methods. No data should be used in the
                     derivation of a criteria until their quality
                     and acceptability had been reviewed by
                     a competent person. Competent people
                     will occasionally disagree, but that is. a
                     fundamental property of subjective
                     decisions.
                       55, Comment—Only published data  .
                     should be used.
                       Response—Peer review is one of
                     many concepts that is better in theory
                     than in practice. Some poor quality data
                     are published and some high quality
                     data are rejected. In addition,
                     publication is not a particularly rapid
                     process. Whether or not data are used
                     should depend on the applicability and
                     quality of the data, not on whether they
                     have been published. Data that  are not
                     published should be made readily
                     available if they are used to derive
                     water quality  criteria.
                       SB. Comment—All static test are
                     unacceptable
                       Response—In general, high quality
                     flow-through acute tests are preferable
                     to high quality static acute tests, but
                     static tests are by no means
                     unacceptable. Few data are available  to
                     show whether static tests consistently
                     produce acute values lower or higher or
                     different than flow-through tests,
                     Whereas degradation, violatilization.
                     and buildup of metabolic products are
                     more likely to be a problem in static
                     tests, operator and mechanical errors
                     are more likely in flow-through  tests.
                     Static acute tests are certainly not
                     unacceptable  for most pollutants, but
                     static chronic tests generally are
                     unacceptable because of changes in the
toxicant concentrations and' the quality
of the dilution water during the test.
  57. Comment—Data obtained, using
test organisms that were previously
exposed to the pollutant should be used.
  Response—Comparisons of results
obtained with unexposed and previously
exposed organisms should indicate:
whether or not acclimation has    .
occurred. Generally, data obtained with
acclimated organisms should not be
used in deriving criteria because
acclimated organisms are the exception
rather than the norm. Rarely, If ever, can
acclimation be depended on to  protect
organisms in a field situation because
concentrations of ten'fluctuate and
motile organisms do not stay in one
location very long. Data obtained with
acclimated organisms may be
acceptable for use in deriving some site-
specific criteria.
  58. Comment—Foreign species should
be used to expand the data base.
  Response—Foreign species may be
representative of indigenous species, but
some of them are quite unusual. Data
obtained with foreign species may give
good indications of indigenous  speceis
that should be used in tests on  some
pollutants and may identify some
potential problems that should  be
investigated.
  59. Comment—-If data for brine shrimp
are not used,  the criteria should not
apply to saline waters.
  Response—Data obtained using brine
shrimp are not used because these
organisms are atypical. Although they
may not be usually sensitive or
insensitive to various pollutants, the
species found in North America and
used for testing only survive in the Great
Salt Lake and in salt ponds near San
Francisco Bay. These two habitats are
unlike any others in the United States. If
criteria were  to be derived specifically
for the Great  Salt Lake or for salt ponds,
then data for brine shrimp should be
used.
  60. Comment—Structure-activity
relationships should not be used unless
proven.
  Response—No provision Is made in
the Guidelines for the use of structure-
activity relationships. Such relationships
may soon be well enough understood
that they can be used in deriving water
quality criteria.
  61. Comment—A  criterion should not
be derived for a pollutant until data are
available for  a broad range of
commercially, recreaiionally, and
ecologically important species. Each
species should be acutely and
chronically tested under a variety of
conditions in a number of different
waters.

-------
                   Federal Register / Vol. 45. No.  231 / Friday. November  28. 1980 / Notices
                                                                       79365
   Response—Except for those people
 who merely want to stop EPA from .
 deriving any water quality criteria, most
 people will admit that there must be
 some reasonable limit as to how much
 information is necessary concerning any
 regulatory action.This is as true for
 deriving water quality criteria, as it is
 for issuing NPDES permits, submitting
 PMNs, registering pesticides, etc. All of
 these regulatory activities deal with
 potentially significant adverse effects on
 aquatic organisms and should take into
 account many of the same possible
 kinds of adverse effects. Therefore, the
 data needs for these various activities
 should probably be somewhat similar,
 but for each regulatory activity the
 minimum data requirements also need
 to take into account the special aspects
 of the program and practical
 considerations. Unrealistic data
 requirements will benefit no one. It is
 not necessary that all questions be
 answered before any action is taken. It
 is only necessary that enough data be
 available to allow reasonable
 confidence that the water quality
' criteria will generally not be too high or
 too low.
   EPA has developed minimum data
 requirements that describe the amounts
 and kinds of information that  should
 usually be available if a criterion is to
 be derived using the Guidelines. When
 the minimum data requirements are
 satisfied, it should usually be possible to
 derive a useful criterion. The
 requirements take into account many
 things such as:
   a. The existence of some species
 which are commerically or
 recreationally important and generally
 sensitive to some broad classes of
 pollutants;
   b. The range of species for which data
 are available;
   c. The cost of obtaining additional
 data and the usefulness of the data; and
   d. The reasonableness of
 extrapolations from one species to
 another within and  between groups.
   The requirements set forth in the
 minimum data base are indeed minimal,
 considering the great varltey of species
 which exist in most aquatic ecosystems.
 However, EPA feels that based on the
 availavble information the routine
 requirement of more data would
 probably not improve criteria  enough to
 justify the additional cost.
   62. Comment—The mimlmum data
 requirements should depend on the
 nature of the pollutant.
   Response—EPA feels that such an
 approach may be feasible some time in
 the future, but would be an unwarranted
 level of sophistication at this time. Fora
 few pollutants, it may be possible to
relax some of the data requirements, but
in general this can only be determined
after enough data are available to
indicate that a special case exists. In
other cases the minimum data may
indicate that additional data are highly
desirable.
  63. Comment—Criteria should not be
derived if enough data are not available.
The alternative procedures which were
proposed should not be used.
  Response—EPA agrees that a
numerical criterion should not be
derived if enough appropriate data are
not available, except in some special
cases. EPA also agrees that the
alternative procedures which were
proposed should not be used to develop
numerical criteria at the present time.
However. EPA feels that when a
numerical criterion is not derived, a
descriptive criterion can be used to
accurately reflect the latest scientific
knowledge.
  64. Comment—The guidelines should
give more guidance on relating a
criterion to a water quality
characteristic.
  Response—More detail on this subject
has been written into the Guidelines.
  65. Comment—If data on the relation
of toxicity and water quality are not
available, no criterion should be
derived.
  Response—The purpose of a criterion
is to present the best available
information, not to ensure that all
desirable.information is available. Any
water quality characteristic may affect
the toxicity of each pollutant to some
degree and it is never going to be
possible to investigate all such
interactions for even a few species and
pollutants. EPA has adopted a minimum
data base requirement for deriving a
criterion, but there must be practical
limits or no criterion will ever be
possible. When the minimum data base
requirements'are satisfied, a criterion
should be derived regardless of
speculation that some unstudied
relationship exist. When enough good
data demonstrate a relation between
toxicity and a water quality
characteristic, an attempt should be
made to use this information in the
derivation of a criterion. A major
purpose of site-specific criteria is to take
into account the effect of local water
quality conditions on toxicity.
  66. Comment—Do not specify the form
that a relationship between toxicity and
water quality must take.
  Response—The Guidelines allow the
use of any set of transformations that fit
the data well. The log-log model  is given
as an example.because it seems  to fit
most of the available data concerning
the relationship between hardness and
toxicity of metals (the only such
relationship for which much quantitative
data are available] reasonably well.
  87. Comment—The toxicity of metals  '
should not be related to "hardness".
  Response—EPA has tried to derive
criteria in a form that will (a) adequately
protect aquatic Organisms and (b) be
practically useful. Hardness is used as
an easily measured surrogate for a
number of interrelated water quality
characteristics, such as pH, alkalinity,
calcium, and magnesium. Various
combinations of these probably affect
individual metals differently, but these
are all reasonably well correlated with
hardness in a wide variety of natural
waters. Some waters, such as those
impacted by acid mine drainage.
obviously are special cases, but they
have special problems of their own.
  68. Comment—Do not extrapolate
slopes for toxicity vs. water quality from
fish to invertebrates or from acute
values to chronic values.
  Response—The Guidelines do not
now assume that the acute slope and the
chronic slope are similar for a pollutant.
On the other hand, there is no reason to
believe that invertebrates are more
similar than are fish and invertebrates.
As explained earlier, the group
"invertebrates" does not consist of a
collection of species that are similar
taxonomically or lexicologically. Some
water quality characteristics apparently
affect the toxicity of the pollutant, rather
than the sensitivity of the organisms. For
these kinds of factors, slopes should be
the same for different  species. Even
factors that affect such things as the
permeability of membranes may
produce similar slopes for a wide
variety of species. If each species must
be treated separately, no  criteria will
ever be possible.
  69. Comment—Relationships based on
only two points should not be used.
  Response—Two  points certainly  do
not provide very much information
about-the shape, slope and position of a
line. However, if other information  or a
reasonable assumption is available
concerning the shape of the line,  two
good data points, spaced at a
reasonable interval, can provide  very
useful information  concerning the slope
and position of the line. Three
appropriately spaced points would
certainly be better, and four points
would be an Ideal minimum.
  70. Comment—Do not combine
relationships that are and are not
statistically significant.
  Response—The Guidelines do now
specify that relationships should be
tested for statistical signficance. A  test
for statistical significance may be one
indication of whether  or not  a slope is

-------
 79386
Federal  Register / Vol. 45. No. 231 / Friday.  November 28, 1980  / Notices
useful, but such a test canr.ot be used
with just two points and does not take
into account such things as the
comparability of the data, the quality of
the test and the range of the
independent variable. A relationship
based on six points may rot be as
significant as i! seems if five of the
points are tightly grouped.
  71. Comment—The Guidelines should
not combine 96-hr LG50 values and 48-hr
EC50 values.
  Response—Both LC50 values and
EC50 values are used to measure acute
toxicity of a substance to aquatic
organisms. In general, an ECSO can be
based on a wide variety of effects, but
the Guidelines specify that the only
effects to be used for deriving criteria
are incomplete shell development.
immcbilization. and loss of equilibrium.
AU of these are certainly drastic effects.
In a field situation these effects
probably often lead to death. Jnst as lie
endpoint may be specific for the species,
so may be the length of the test. The
generally accepted length of an acute
test with daphnids Is 48 hours, whereas
for most species of fish, it is 96 hours.
Thus the Guidelines use both 48-hr ECSO
values and 98-hr LC50 values because
they are the widely accepted durations
and endpoints used to measure acute
toxicity to specific species.
  72. Comment—Shell deposition tests
are chronic tests and should not be
equated with lethality tests.
  Response—"Acute" implies "short"
not "death". Many acute toxicity tests
do use death for the effect, but many
also use non-lethal effects. The shell
deposition test is one of many non-lethal
acute tests and is generally accepted as
a short test compared to the average life
span of oysters.
  73. Comment—Adjustment factors
should not be used to adjust for the
length of the test, the technique, and
unmeasured concentrations.
  Response—All three kinds of
adjustment factors have been deleted
from the Guidelines. The factor for the
length of the test was found to be
unnecessary because most tests had
been conducted for the standard times
usually specified for the individual
species. Thus the Guidelines now
specify that only data from tests
conducted for the time specified for the
species should be used to calculate the
Final Acute Value.
  EPA has found that on the average
flow-through acute tests give results
slightly lower than do static tests, but
the relationship does not seem to be too
consistent and may vary from species to
species for some pollutants. In addition,
on the average results based on
measured concerntrations do not seem
                     to be much different from those based
                     on unmeasured concentrations.
                        However, the results of flow-through
                     tests based on measured concentrations
                     are generally accepted as being better
                     measures of acute toxicity than the
                     results of flow-through tests based on
                     unmeasured concentrations or the
                     results of any static or renewal
                     tests.Thereforc. whenever the results of
                     flow-through acute tests in which the
                     concentrations were measured are
                     available, the results of all other kinds
                     of acute tests with that species and
                     pollutant are not used in the calculation
                     of the species mean acute value.
                        74. Comment—Species sensitivity
                     factors should be pollutant-specific; and
                     average factor should  not be calculated
                     for a variety of substances.
                        Response—EPA agrees. The
                     requirement for acute values for at least
                     eight different species was developed in
                     part to allow for a reasonably good
                     calculation of a mean acute value and a
                     species sensitivity factor for each
                     individual pollutant. A better way of
                     using the acute values for the individual
                     species has been developed, but no
                     extrapolations are made from one
                     pollutant to another.
                       75. Comment—The distribution of
                     species mean acute values for a
                     pollutant will be truncated if the species
                     cannot be killed or affected by
                     concentrations above solubility.
                       Response—Some species are so
                     resistant to some pollutants that they
                     cannot be killed or affected in acute
                     tests even by concentrations which are
                     much above solubility. Such "greater
                     than" values cannot be used in the
                     calculation of means and variances for
                     pollutants. When the "greater than"
                     values are for insensitive species and
                     are at or above solubility, the values can
                     be used in the calculation of the Final
                     Acute Value by adjusting the cumulative
                     proportions for all the  spelces with
                     quantitative values. The shape of the
                     curve at the high end cannot be
                     determined, but the Final Acute Value is
                     more dependent on the species mean
                     acute values and the cumulative
                     probabilities at the low'end.
                       76. Comment—Early life-stage tests
                     with fish should be used
                     interchangeably with life-cycle and
                     partial life-cycle tests  with fish.
                       Response—EPA agrees that early life-
                     stage tests with fish generally give about
                     the same results as comparable life-
                     cycle and partial life-cycle testa.
                     However,Nbecause the shorter test is
                     merely a predictor of the longer tests,
                     whenever both kinds of results are
                     available, the results of life-cycle and
                     partial life-cycle tests  should be used
instead of the results of oariv life-stage
tests.
  77. Comment—Appropriate measures
of chronic toxicity and appropriate
lengths of exposure should be defined.
  Response—The descriptions of
appropriate chronic tests have been
clarified.
  78. Comment—The factor of 0.44
should not be used.
  Response—It is not now used.
  79. Comment—The Final Chronic
Value should not be lower than the
lowest measured species chronic value,
even if chronic data are not available
for sensitive species.
  Response—Aquatic ecosystems
cannot be protected from chronic
toxicity by protecting only the
insensitive species from chronic toxicity.
In the past both arbitrary and
experimentally determined application
factors have been used to relate acute
and chronic toxicity. For a variety of
reasons the Guidelines do not use an
application factor, but instead  use the
acute-chronic ratio, which is similar to
the inverse of an application factor.
Thus the acute-chronic ratio should
normally be greater than one. The acute-
chronic ratio is to be used with
invertebrates as well as fish and is to be
an experimentally determined  value for
each individual pollutant The  acute-
chronic ratio should also avoid the
confusion as to whether a large
application factor Is one that is close to
unity or one that has a denominator that
is much larger than  the numerator. The
acute-chronic ratio is calculated by
dividing the appropriate measure of
acute toxicity for the species (as
specified in the Guidelines) by the
appropriate measure of chronic toxicity
for the same species (as specified in the
Guidelines).
  Some people have confused
application factors and safety  factors
and use of the term "acute-chronic
ratio" should help avoid this problem.
Acute-chronic ratios are a way of
estimating the chronic sensitivity of a
species for which no chronic toxicity
data are available. Safety factors would
provide an extra margin of safety
beyond the sensitivity of the species.
Safety or uncertainty factors are
intended to reduce the possibility of
underprotection, whereas acute-chronic
ratios are intended  to estimate the
actual chronic sensitivity of the species
to the pollutant. This estimate  is just as
likely to be too high as it is to be too
low. A mean acute-chronic ratio will in
fact be too high for half the species and
too low for the other half.
  When three or more acute-chronic
ratios have been determined for a
pollutant with both fish and

-------
                   Federal Register / Vol. 45. No. 231 / Friday.  November 28. 1980 / Notices
                                                                      79367
  -.vertebrates, three patterns have been
  ibserved when the individual species
  ire listed in order of their species mean
  cute values:
   a. The ratios randomly differ by a
  actor'of ten or more.
   b. The ratio appears to be about the
  ame (within a factor of ten) for all
  pecies.      .
   c. Species with higher acute values
  ilso have higher acute-chronic ratios.
   The available data indicate that fish
  ind Invertebrates do not consistently
  lave different acute-chronic ratios and
  hat for some pollutants freshwater and
  saltwater species have similar acute-
  :hronie ratios.
   80. Comment—No application factor
  hould be used unless it is specific for
  he pollutant, species, and water.
   Response—There la no point in using
  in application factor or acute-chronic
  atio or any concept if It does not allow
  •ome generalization or extrapolation
  rom one species  to another or from one
  vater to another. Not allowing'any
  jeneralizations or extrapolations would
  equire that much data be generated for
  >ach species and  each pollutant in each
  valer in which a criterion is necessary.
  When enough supporting data, are
  ivailable, extrapolations using such
  hings as acute-chronic ratios are cost-
  iffective and scientifically sound.
   81. Comment—Additional
  levelopment of methodology for toxicity
  ests with aquatic plants is needed:
   Response—This is most certainly true.
 vluch other research also Is needed, and
  ;enerally is considered higher priority.
 ZPA. hopes that someday all of the
  iddltional research that needs to be
  lone will be done. Few pollutants seem
  o affect aquatic plants at
 -.oncentrations which do not chronically
 iffect aquatic animals, and it is hoped
  hat this is not an artifact of the test
 nethods currently used.
 :  82. Comment—Data on toxicity to
 ilants should not be used for deriving
 •jiteria because plants are more site-
 specific than animals.
   Response—Numerous species of
 itants, especially algae, exist in most
 lodies of water. On the other hand, EPA
 uiows of no data! to support the
 :ontentioh that the sensitivities of
 iquatlc plants are any more site-specific
  han those of aquatic animals, or that
:  he range of sensitivities between plants
  s as great as that for animals. One
 species may or may not be
  epresentative of other species. After the
 nethodology for toxicity tests with
 iquatic plants is better developed, tests
  vith a wider variety of species would
 :ertainly be desirable.
   83. Comment—The Final Plant Value
 ihould not be the  lowest available plant
value based on measured
concentrations.
  Response—EPA adopted the
procedure described in the Guidelines
for obtaining the Final Plant Value for
several reasons including:
  a. The methodology for toxicity tests
with aquatic plants is not well
developed.
  b. For only a few pollutants have
toxicity tests been conducted with more
than a very few species of plants.
  c. Little Is known about the range of
sensitivities of various species of
aquatic plants.
  d. Based on available data, almost no
pollutants are toxic to aquatic plants at
the lowest concentrations which are
chronically toxic to aquatic animals or
cause unacceptable residues.
  84. Comment—Residue accumulation
in any part of an aquatic ecosystem
should be prevented as much as
possible.
  Response—Accumulation of residues
in aquatic organisms only becomes a
problem if the concentration of residue
is high enough to adversely affect either
(a) the organism itself, (b] a consumer of
the organism, or (c) the'marketability of
the organism. Adverse effects on the
aquatic organism itself will be detected
in acute and chronic toxicity tests. The
use of FDA action levels and chronic
feeding studies with wildlife  are
designed to protect the uses and
consumers of aquatic organisms.
  85. Comment—Bioconcentration
factors (BCFs) derived from field data
should not be used.
  Response—EPA feels that BCFs
derived from adequate data, whether
they be laboratory data or field data,
should be used. More data are
necessary to document a BCF from a
field exposure than a laboratory
exposure, as specified in the Guidelines,
but if enough data are available, field
BCFs should be used.
  86. Comment—Kinetically derived  •
bioconcentration factors (BCFs) should
be used.
  Response—Kinetically derived BCFs
should be used if the bioconcentration
test lasted long enough, i.e., to .apparent
steady-state, to verify that the model
(assumptions) used in the calculations
actually fits the data for the individual
pollutant.
  87. Comment—Bioconcentration
factors (BCFs) should not be estimated
from octanol-water partition
coefficients.
  Response—The available data seem
to indicate a reasonably good
relationship for lipid-soluble substances
between steady-state BCFs and octanol-
water partition coefficients. BCFs
estimated from partition coefficients are
not used in the Guidelines because
measured BCFs are available for all
pollutants for which a maximum
permissible tissue concentration is
available.
  88. Comment—Bioconcentration
factors (BCFs) are dependent on
temperature, food, salinity, stress, and
other things.
  Response—Many things such as these
probably do affect BCFs. Until data are
available to show that such effects are
important and are not species-specific.
little needs to be, or can be. done to take
such factors into account when deriving
water quality criteria.
  89. Comment—Bioconcentration
factors (BCFs) should be based only on
tissues that are actually eaten.
  Response—Although people usually
only eat muscle tissue of fish, wildlife
usually eat the whole body of fish. The
tissues used in the determination of
BCFs must be appropriate to the kind of
consumer organism or regulatory action.
On the other hand, since the BCF for a
lipid-soluble substance seems to be
proportional to percent lipids;
extrapolations can be made on the basis
of percent lipids regardless of the tissue.
  90. Comment—Chronic toxicity tests
with rats and mice should not be used
as representative of tests on mammalian
wildlife.
  Response—Because results of tests on
a variety of species are extrapolated to
man, it should be just as reasonable to
extrapolate from one mammalian
species to another mammalian species
within certain limits. However, such
extrapolations are not now used in the
Guidelines; only the results of chronic
toxicity tests with wildlife are used to
protect wildlife consumers of aquatic
life.
  91. Comment—Information concerning
bioconcentration should only be used if
such information is used to protect
aquatic organisms, not to protect the
marketability of aquatic organisms.
  Response—Protection of aquatic
organisms must include not only the
protection of the existence of aquatic
organisms, but also protection of the
common uses of aquatic organisms.
Commercially important aquatic
organisms cannot be considered
adequately protected if they cannot be
sold. The Guidelines do not use any
data pertaining to safety to humans in
an attempt  to protect human consumers
of aquatic organisms. Instead, the
Guidelines  merely attempt to ensure
that residues in aquatic organisms do '
not exceed FDA action levels so that the
uses of commercially and recreationally
important species are not restricted by
the Food and Drug Administration.

-------
 79368
Federal Register  /  Vol.  45, No.  231 / Friday. November 28, 1980 / Notices
   92. Comment—A Final Residue Value
 calculated from an FDA action level is
 actually a concentration that will result
 in the average concentration in some
 species being at the FDA action level.
   Response—This is a good point. A
 similar situation exists when the
 calculation is based on a concentration
 which caused an adverse effect in a
 chronic wildlife feeding study. In all
 such cases, the Final Residue Value
 should be lower, but EPA knows of no
 non-arbitrary way to determine how
 much lower the value should be.
  93. Comment—The FDA action levels
 for finished animal feed should not be
 used,   .
  Response—They are not now used.
  94. Comment—Flavor impairment
 should not be used to derive water
 quality criteria for aquatic life.
  Response—Many of the commercially
 and recreatlonally important aquatic
 orgnaisms are consumted by people. If
 the flavor is significantly impaired, the
 use of these species will be adversely
 affected. Flavor impairment should  be
 considered an effect that can adversely
 affect the use of aquatic organisms.
  95. Comment—The Instructions for
 using the other data are not very
 detailed and are not mathematical.
  Response—EPA has tried to include
 as much detail in the instructions for
 using the other data as are currently
 justified. Extensive detail and
 mathematical treatment are not deemed
 realistic at this time because so little
 information is available concerning the
 various kinds of other data.
  90. Comment—The final review of the
 criteria should allow revision up or
 down based on sound scientific
 evidence.
  Response—The  Guidelines always
 have allowed revision up or down, but
 this is now stated  explicitly in the
 Guidelines,
  97. Comment—Some bodies of water,
 such as some USGS benchmark streams
and the Houston ship channel, contain
 concentrations above the criteria for
some pollutants and still contain aquatic
 communities that are diverse, healthy,
and productive. Such information should
 be used in the review of the criteria
because it indicates that  some criteria
are too low.
  Response—Rarely are  there enough
 data available to accurately identify the
 concentrations of pollutants to which
 aquatic organisms in bodies of water are
actually exposed. The sampling scheme
should provide a good estimate of the
 mean and variance of the concentration;
a few grab or composite samples cannot
 provide enough information to
characterize the concentrations of
 pollutants  in most bodies of water. The
                    concentrations vary not only with time
                    but also with location at each time, so
                    the samples must be taken where the
                    organisms of interest are located at that
                    time.
                      A more serious problem concerns the
                    definition of an acceptable aquatic
                    ecosystem. How does one determine if
                    an aquatic ecosystem Is healthy or
                    productive? If a diverse system is, by
                    definition, healthy, is it also, by
                    definition, productive? What is the
                    minimum acceptable diversity? What is
                    the minimum acceptable productivity?
                    Should the acceptable levels of diversity
                    and productivity be site-specific? Is a
                    body of water acceptable just because
                    no dead fish are observed. How many
                    pounds of trout should a trout stream
                    produce each year to be considered
                    healthy and productive? How does one
                    treat motile species that may avoid
                    some periodic increases in pollution
                    levels? Is an aquatic ecosystem healthy
                    and productive if the normally edible
                    portion of a consumed species tastes
                    bad or contains excessive residues?
                    Questions such as these indicate the
                    difficulty of quantitatively judging the
                    quality of aquatic ecosystems on the
                    basis of their acceptability or usefulness
                    to man or on any other basis. Although
                    judging bodies of water would be a
                    difficult job, it certainly could be done
                    by a competent group of trained
                    professionals. The point is that it is not
                    as easy a job as some people would like
                    to think. There are also people who feel
                    that various pristine bodies of water
                    should be managed because they are not
                    as productive as they could be.
                      As mentioned earlier, the criteria
                    documents derive criteria which may be
                    too high or too low for some specific
                    bodies of water. With appropriate
                    modifications the Guidelines can be
                    used to derive criteria for any specific
                    body of water or geographic area. In
                    addition, it is certainly possible that one
                    or more factors which affect the toxlcity
                    of one or more pollutants may not have
                    been studied very throughly or even
                    identified yet. The criteria are based on
                    the best available  information and the
                    state-of-the-art of aquatic toxicology,
                    but it IB always possible that something
                    important has not been adequately
                    studied by regulators, discharges or
                    academia.
                    Appendix E.—Responses to Public
                    Comments on the Human Health Effects
                    Methodology for Deriving Ambient
                    Water Quality Criteria
                    /. Introduction
                      On March 15.1979, the U.S.
                    Environmental Protection Agency (EPA)
                    announced the availability for public
comment of the proposed
for the derivation of ambient
quality criteria for the protection of
human health. The public conitnenti
were resolved in three phases.
  First, comments relating to policy
issues were resolved in an initial
screening/disposition by Agency
personnel.' Second, a peer review
workshop was conducted and Involved
•Agency personnel, contractors, and
recognized scientists. The group
evaluated all issues pertaining to the
derivation of criteria for non-
carcinogens, and third, a similar
workshop was held to review all issues
relating to the derivation of criteria for
carcinogens.
  The following report presents the
resolutions of the public comments by
the EPA after considering the advice of
the meeting attendees. While the EPA
greatly appreciates the contribution of
these individuals  and acknowledges
their substantial assistance in resolving
many difficult questions, the EPA
accepts full responsibility for the
positions outlined in  this document.
(Note: Comments addressing similar
issues were appropriately compiled and
summarized under each issue.)

Comments Resolved in Initial Screening

Issue!                  JJ.,--
  Comment summary: The water quality
criteria documents should provide
information and/or guidelines for
deriving standards from criteria.
  Response: The water quality  criteria
documents contain information which
will be useful in developing standards
(e.g., current levels of exposure).
However, in developing standards,
many additional factors not directly
related to criteria must be considered. It
would be more appropriate to compile
and to analyze this information as part
of the standard-setting process rather
than to Include it  In the criteria
documents. Guidelines will be issued
separately since the development of the
standard includes use designation with
a commensurate criteria value.
Issue 2
  Comment summary: Water quality
criteria should consider or be limited by
technological achlevability, cost/benefit
analysis, limits of detection, and
environmental fate,    .
  Response: The distinction between
criteria and standards must be
recognized. For non:carcinogens,
ambient water quality criteria are
estimates of concentrations in water
which will not result in either adverse
human health effects (criteria based on
toxicity)  or unplesant taste or.odor

-------
                   Federal Register  /  Vol. 45. No. 231 / Friday. November 28. 1980  / Notices             79369
(organoleptic criteria). For carcinogens.
criteria are estimates of concentrations
of individual compounds in water which
will result in specified increases in the
lifetime risk of developing cancer. By
definition, these criteria exclude
considerations of technological
achievability, .cost/benefit analysis,
limits of detection, and environmental
fate, as appropriate within the authority
of The Clean Water Act (33 U.S.C.
1314(a)]. These factors are more
properly considered in the standard-
setting process.
Issue 3
  Comment summary: The validity of a
single criteria for all bodies of water is
questionable. Criteria should be site
specific an/or use specific.
  Response: In the standard-setting
process, criteria may be modified based
upon site specific or use specific
considerations.

Issue 4
  Comment summary: Even if there is
insufficient data, some criteria must still
be developed for "highly hazardous
compounds."
  Response: If there is sufficient
information to indicate that a compound
is "highly hazardous," there should be
sufficient information to derive a
criteria. Conversely, if Insufficient data
are available, by definition no criteria
can be derived.
Issue 5
  Comment summary: Criteria should
be derived only for persistent
compounds or for compounds which
present a clear hazard to humans.
  Response: Criteria can be derived for
any compound on which sufficient
information is available. By definition.
criteria are independent of persistence
or current levels of exposure.
Issue 6
  Comment summary: Criteria should
be developed to protect terrestrial
wildlife as well as humans and aquatic
organisms.
  Response: Because of the greet
number of diverse wildlife species and
differences in their habitat, diet, and
behavior, it is unlikely that a single
criteria could be developed to protect all
wildlife species from a given
contaminant. The EPA is currently
assessing possible approaches to
developing  a valid methodology for
deriving wildlife criteria. Until a specific
wildlife criteria methodology is
developed,  the proposed aquatic life and
luman health effects criteria should
serve as interim levels for the protection
>'• -vldlife.
Issue 7

  Comment summary: Criteria should
be derived by an independent scientific
panel and not by the EPA.
  Response: The EPA has a legislative
mandate to derive ambient water
quality criteria and must accept the final
responsibility for this process. However,
the EPA has solicited the advice of
many independent scientists in this
effort. It should be noted that the
consensus of the peer review
committees has been considered and
generally followed by the EPA.
Nonetheless, the responsibility for the
criteria rests solely with the Agency.

Issue 8

  Comment summary: The ambient
water quality criteria are not sufficiently
protective of special groups at risk.
  Response: In most cases, each
document contains a specific section on
special groups at risk. This is intended
to serve as a notice to individuals or
agencies using the criteria; that the
derived criteria may not be sufficiently
protective in all applications. If
sufficient data are available,
information in the section on special
groups at risk could be used to modify
the criteria during the standard-setting
process.

Issue 9

  Comment summary: Comments
express concern with the failure of the
criteria to specifically address possible
toxicant interactions.
  Response: The importance of toxicant
interactions in the environment cannot
be disregarded. Each document attempts
to summarize the available data on such
interactions. However, since the
composition of toxicants is likely to vary
substantially in different area?, a
general approach modifying criteria
based upon toxicant interactions is not
available at this time. Further, the
limitations of valid approaches for
dealing with interactions in multi-
toxicant mixtures should be recognized.

Issue 10

  Comment summary: Because of the
uncertainties involved in deriving
criteria, the criteria should be limited to
only one significant figure.
  Response: The number of significant
figures used to express the criteria is an
admittedly arbitrary decision. The EPA
recognizes the inexactitude of these
numbers.
///. Comments on Non-Carcinogens
A. Criteria for Chemical Classes

Issue 1
  Comment summary: Two basic
approaches were taken in the
documents on chemical classes when
sufficient data were not available on all
members in a class:
  (a] Criteria were derived for
individual chemicals on which sufficient
data were available and no criteria were
recommended for other chemicals in the
class.
  (b) A criteria was derived for all or
some chemicals in  the class based on
toxicity data on one or a few members
of the class.
  Alternative "a" can be  criticized for
"allowing" contamination .by "probably
hazardous compounds" (reasoning by
chemical analogy). Alternative "b" can
be criticized for applying a general
criteria to a specific compound for
which data are not available.
  What guidelines  with justifications
can be given for selecting either
alternative? What other alternatives
might be considered?
  Response: The initial methodology did
not adequately address the problems
associated with deriving class criteria.
The following section has been added to
the methodology and serves as a useful
guide in the criteria derivation process.
  A chemical class is broadly defined as
any group of compounds which are
considered in a single risk assessment
document In criteria derivation, isomers
ere regarded as a chemical class rather
than as a single compound. A class
criteria is an estimate of risk/safety
which applies to more than one member
of a class, and involves varying degrees
of extrapolation from available data on
some members of the class to other
class members on which sufficient data
are not available to derive a compound-
specific criteria (i.e., a criteria based on
data solely on the specific chemical  for
which the criteria is derived).
  A class criteria usually applies to
each member within the class rather
than to the sum of the compounds within
the class. While the potential hazards of
multiple toxicant exposure are not to be
minimized, a criteria, by definition, most
often applies to an  individual
compound. Exceptions may be made of
complex mixtures which are produced,
released, and lexicologically tested  as
mixtures (e.g., toxaphene and PCBs). For
such exceptions, some attempt should
be made to assess  the effects of
environmental partitioning different
patterns of environmental transport and
degradation on the validity of the
criteria. If these effects cannot be

-------
 79370
Federal  Register / Vol. 45. No. 231  / Friday. November  28. 1980 / Notices
assessed, an appropriate statement of
uncertainty should accompany the
criteria.
  Because relatively minor structural
changes within a class of compounds
can have pronounced effects on their
biological activities, class criteria should
be avoided. Whenever sufficient
toxicologic data are available on a
chemical within a class, a compound
specific criteria for that chemical should
be developed. Nonetheless, for some
chemical classes, scientific judgment
may suggest a sufficient degree of
similarity among chemicals within a
class to justify a class criteria applicable
to some or all members within a class.
Such a judgment should be influenced
by a perceived risk to  the human
population if a class criteria was not
derived.
  The development of a class criteria
should take into consideration the
following:
  (a) A detailed review of  the chemical
and physical properties of  chemicals
within the group should be available. A
close relationship within the class with
respect to chemical activity would
suggest a similar potential  to reach
common biological sites within tissues.
Likewise, similar lipid solubilities would
suggest the possibility of comparable
absorption and tissue distribution.
  (b) The amount of qualitative and
quantitative data for chemicals within
the group should be examined.
Obviously adequate toxicologies! data
on a number of compounds within a
group would provide a more reasonable
basis for extrapolation than minimal
data on one or two chemicals within a
group.
  (c) Similarities In the nature of the
lexicological response to chemicals in
the class provides additional support for
the prediction that the response to other
members of the class may be similar. In
contrast, where the biological response
has been shown to differ markedly on a
qualitative and quantitative basis for
chemicals within a class, extrapolation
of a criteria to other members of that
class may not be appropriate.
  (d) Additional support for the validity
of extrapolation of a criteria to other
members of a class could be provided
by evidence of similar metabolic and
pharmacokinetic data, if available, for
some members of the class.
  Based on the above considerations, it
may be reasonable to divide a chemical
class into various subclasses. Such
divisions could be based on biological
endpoints (e.g.. carcinogens/non-
carcinogens), potency, and/or
sufficiency of data (e.g., a criteria for
some members of a class but no
criterion for others). While no a priori
                     limits can be placed on the extent of
                     subclasslfication. each must be
                     explicitly justified by the available data.
                       Class criteria, if properly derived and
                     supported, can constitute valid scientific
                     assessments of potential risk/safety and .
                     can be used in establishing appropriate
                     standards. Conversely, the development
                     of a class criteria from an insufficient
                     data base can lead to serious errors in
                     underestimating or overestimating risk/
                     safety and should be rigorously avoided.
                     Although scientific judgment has a
                     proper if not totally explicable role in
                     the development of class criteria, such
                     criteria will be useful and defensible
                     only if they are based on adequate data
                     and scientific reasoning rather than
                     intuition. The lack of data on
                     dissimilarity cannot be used as the basis
                     of a class criteria. Further, the definition
                     of sufficient data on similarities in
                     physical, chemical, pharmacokinetic, or
                     toxicologic properties to justify a class
                     criteria may vary remarkably depending
                     on the degree of superficial structural
                     similarity and the gravity of the
                     perceived risk. Consequently, it is
                     imperative that the criterion derivation
                     section of each document in which a
                     class criterion is recommended
                     explicitly address each of the key issues
                     discussed above and define, as clearly
                     as possible, the limitations of the
                     proposed criteria and the type of data
                     necessary to generate a compound-
                     specific criterion.
                       Class criteria should be corrected
                     when sufficient data become available
                     to derive a compound-specific criterion
                     that protects against the biological effect
                     of primary concern. The availability of a
                     good subchroni study would not result
                     necessarily in the abandonment of a
                     class criteria based upon potential
                     carcinogeniciry.
                       The Inability to derive a valid class
                     criteria does not and should not
                     preclude regulation of a compound or
                     group of compounds based upon
                     concern for potential human health
                     effects. The failure to recommend a
                     criterion is simply a statement that the
                     degree of concern cannot be quantified
                     from the available data and risk
                     assessment methodology.
                     Issue 2

                       Comment summary: To  wrjat extent
                     can "guilt by association" be used to
                     derive a cancer-based'criteria for a
                     compound which has been tested for'
                     carcinogenic!ty with negative results
                     [e.g., bis(2-chloroisopropyl) ether in the
                     Chloroalkyl Ethers Ambient Water
                     Quality Criteria Document].
                       Response: As stated in the response to
                     Issue 1. "guilt by association" is only an
extremely limited role in criteria
derivation process.

B. Organoleptic Criteria  .

Issue 3

  Comment summary: Whenever
organoleptic criteria are derived,
corresponding toxicity based criteria
should be derived if possible.
  Response: The Agency agrees. Since
organoleptic criteria are not based on
toxicologic information and have no
direct relationship to potential adverse
human health effects, both organoleptic
and toxicity based criteria are provided
whenever possible.

Issue 4

  Comment summary: The quality of
organoleptic criteria should be assessed
in terms of experimental design and
statistical analysis.   .
  Response: The revised methodology
recognizes the limitations of most
organoleptic data:
  With very few exceptions, the
publications which report taste and odor
thresholds  are cryptic in their
descriptions of test methodologies,
number of subjects tested,
concentration/response relationships,
and sensory characteristics at specific
concentrations above the threshold.
Thus the quality of the data is usually
worse than the lexicological data used
for the setting of other criteria.
Consequently, a clear critical evaluator
of the available data on a compound's
organoleptic characteristics should
appear in the criteria document.

Issue 5

  Comment summary: Criteria based  on
organoleptic properties should not be
considered equal to criteria based on
toxicologic effects.
  Response: The revised methodology.
makes a clear distinction between
organoleptic and toxicity based criteria.
The use of the criteria in the regulatory
process should reflect an  appreciation of
this distinction.

C. Naturally Occurring Compounds

Issue 6

  Comment summary: Background
levels should be defined in terms of the
quality of the data base and
geographical/seasonal variations.
  Response: The documents summarize
data on background levels of naturally
occurring compounds and include
information on seasonal and/or
geographical variation when available.

-------
                    Federal Register / Vol.  45. No. 231 / Friday. November 28. 1980  /  Notices
                                                                       79371
  Issue 7
    Comment summary: A distinction
  should be made between natural and
  anthropogenic background.
    Response: An attempt is made, with
  extfeme difficulty, in the exposure
  section of the documents to differentiate
  between natural and anthropogenic
  background. However, background
  levels cannot be used directly to modify
  the criteria. By definition, criteria should
  not consider current levels of exposure
  but are estimates of safe level or
  incremental risk level exposures.
  Background levels, both natural and
  anthropogenic, should be considered if
  the criteria are used to promulgate
  standards.
  Issues
   Comment summary: What is the
  minimum data base needed to define a
  compound as essential?
   Response: As indicated In the revised
  methodology, elements will be accepted
  as essential if the National Academy of
  Sciences (NAS) Food and Nutrition
  Board or a comparably qualified group
  declares them as such. Elements not yet
  determined to be essential, but for
  which supportive data on "essentiality"
  exists, were recommended to be
  reviewed by a joint EPA/NAS
  committee.
  Issue 9               .
   Comment summary: How can
  essentiality be used to modify a criteria?
   Response: The. following additions
  have been made to the revised
  methodology in response to this
 question:
   In order to be useful in modifying
 toxicity/carcinogenicity based criteria,
 essentiality must be quantified either as
 a recommended daily allowance (RDAJ
 or minimum daily requirement (MDR).
 These levels must be compared to
 estimated daily doses associated with
 the adverse effect of primary concern.
 The difference between the RDA or
MDR and the daily, doses causing a
 specified risk level for carcinogens or
 acceptable daily Intake (ADI) for non-
 carcinogens defines the "window" of
 daily doses from which the criteria
 should be derived.
   Because errors are inherent in
 defining both essential and maximum
 tolerable levels, the criteria should be
 derived from dose  levels near the center
 of such a dose range. The decision to
 use either the MDR or RDA will be
 guided by the size of the window and
 the quality of the essentiality and
 toxicity estimates.
   The modification of criteria by
 consideration ofessentiality must
 include all routes of exposure. If water
 is a significant source of the MDR or
 RDA, the criteria must allow for
 attainment of essential intake.
 Conversely, even when essentiality may
 be attained from non-water sources,
 standard criteria derivation methods
 may be adjusted if the derived criterion
 represents a small fraction of the ADI or
 MDR. On a case-by-case basis, .the
 modification in the use of the guidelines
 may include the use of different safety
 factors for non-carcinogens or other
 modifications which can be explicitly
 justified.
 D.UseofNOAELs/NOELs

 Issue 10
   Comment summary: NOELs and
 related effect terms should be defined
 more clearly in the methodology.
   Response: In the revised methodology,
 the following additions have been made
 to clarify the use of these terms:
   In developing guidelines for deriving
 criteria based on non-cardnogenic
 responses, five types of response levels
 are considered:
 NOEL—No-Observed-Effect-Level
 LOEL—Lowest-Observed-Effect-Level
 NOAEL—No-Observed-Adverse-Eflect-Level
 LOAEL—Lowest-Observed-Adverse-Effect-
    Level
 FEL—Frank-Effect-Level
 In the above terms, adverse effects are
 defined as any effect resulting in
 functional impairment and/or
' pathological lesions that may affect the
 performance of the whole organism, or
 which contributes to a reduced ability to
 respond to an additional challenge. The
 word lowest refers to the incidence of
 the effect in the tested population. It
 should be noted that LOELs, NOAELs.
 and LOAELs refer to exposure levels or
 dosage zones which are experimentally
 defined by upper and lower exposure
 levels. NOELs and FELs, however, are
 not defined at the lower and upper
 exposure levels, respectively.
 Issue 11   ,        ...-..'   : ,.
   Comment-summary: Considerations of
 experimental design should be more
 explicitly/quantitatively considered in
 the criteria derivation process.
   Response.-Tbe development of a rigid
 system for considering experimental
 design in criteria derivation would limit
 the use of scientific judgment The ,
 section of the methodology dealing with
 the derivation of toxicity based  criteria
 has been extensively revised to allow
 for the maximum use of scientific
 judgment in selecting safety factors
 based on both  the quality of the
 individual study and the weight of the
 supporting scientific data.
 E. Safety or Uncertainty Factors

 Issue 12
   Comment summary: Can the
 guidelines for applying safety factors be
 clarified or developed in greater detail
 to minimize inconsistencies without
 impairing scientific judgment?
   Response: The following additions
 have been included in the methodology
 to allow for the use of greater judgment
 in the application of safety factors,
 while also requiring more, explicit
 justification for the use of any
 uncertainty, facton
   The justifications for the various
 safety factors can become very
 restrictive if they are not employed with
 care and judgment. This is the case
 especially In those instances where the
 data do not completely fulfill the
 conditions for one category of
 uncertainty factor and appear to be
 intermediate between two categories.
 Given the uncertainties in the entire  ,
 process, it is more appropriate to set the
 operative uncertainty factor at some
 intermediate-value on a logarithmic
 scale (e.g., 32, being halfway between 10
 and 100 on a logarithmic scale). If
 intermediate values for uncertainty
 factors are more representative of actual
 conditions, then they are used.
   In the  selection of the uncertainty
 factor approach, "no indication of
-carcinogenicity" is interpreted as the
 absence of carcinogenic data from
 animal studies or human epidemiology.
 Short-term carcinogenicity screening
 tests are considered in the criteria
 documents, and are used in the   .
 derivation of numerical criteria and  are
 used to rule out the uncertainty factor
 approach.
   Because of the high degree of
 judgment involved in the selection of a
 safety factor, the criteria derivation
 section of each document must provide
 a detailed'discussion and justification
 for both  the selection of the safety factor
 and the data for which it is applied. This
 discussion should reflect a critical
 review of the total data base. Factors to
 be considered Include: number of ,.
 animals  tested, parameters tested,
 species tested, quality of controls, dose
 levels, route, dosing schedules, etc. An
 effort should be made to differentiate
 between coherent results which form a
 lexicologically valid data base and data
 which may be spurious in nature.
 Issue 13
   Comment summary: What, if any,
 safety factor should be used when
 deriving criteria from a threshold limit
 value (TLV).
   Response: The safety factor used
 when deriving criteria from a TLV must

-------
 79372
Federal Register /  Vol.  45, No.  231 / Friday, November 28, 1980 /  Notices
 depend on the quality of the data base
 on which the TLV is based,
 considerations of uncertainties involved
 in extrapolating data from inhalation to
 oral exposures, and the quality of the
 additional supporting data.
 F. Related NOAEL, Issues

 Issue 14
   Comment summary: Can/should
 concentration response curves
 representing a "full range of effects" be
 used in deriving criteria?
  Response: No available system for
 utilizing concentration response curves
 in representing B full  range of effects for
 deriving criteria has been developed". If
 such a system does become available, it
 will  be assessed by the Agency,
 Issue IS
  Comment summary: When more than
 ODS method Is available to derive a non-
 carcinogen criteria (e.g., 2-year chronic,
 80-day, TLV), can guidelines be given for
 selecting the most appropriate method?
  Response: As Indicated in the revised
 methodology, criteria can be based on
 several different types of data (e,g,,
 studies on humans or experimental
 animals, subchronic or chronic exposure
 periods, oral or Inhalation exposure
 routes, TLVs or similar standards).
 Specific guidelines for selecting a
 particular study or approach have not
 been recommended because of the many
 Judgments] factors which are involved.
 As indicated in the methodology,' the
 criteria derivation section must
 specifically state the reasons for
 selecting the approach and study used
 to derive the criteria.
 Issue IB
  Comment summary: The approach
 used to derive criteria for non-
 carcinogen^ Siay not adequately address
 the ques;;, •• i of whether children are at
 greater n*k tlmrj adults.
  /ZespcM'.ierWhen specific data are
 available on women or children as
groups at increased risk, it should be-
 stated in the document and discussed in
 the crileiia derivation section, but
 should be used to modify the criteria
 only if sufficient specific data are
 available. This is a highly Judgmental
 decision which must be made on an
 individual case.
Issue 17
  Comment summary: Criteria based on
carcinogenic effects might not be
 adequate to protect humans from
mutagenic, teratogenic, or other toxic
 effects.
  Response: With very few exceptions,
 criteria based on carcinogenicity are
probably protective for other toxic
                     effects. However, alternative criteria
                     can be derived based qn non-
                     carcinogenic effects on a case-by-case
                     basis if there is'any doubt of the level of
                     protection offered by the cancer based
                     criteria.
                     C. Alternative Approaches to the
                     Development of Criteria for Non-
                     Carcinogens

                     Issue 18
                      Comment summary: Is there a
                     reasonable way to use multiple NOEL/
                     NOAELs to derive criteria?
                      Response: The revised methodology
                     clearly Indicates that all toxicity must
                     be considered in deriving criteria and
                     multiple NOELS/NOAELs are used. A
                     detailed mathematical approach using
                     multiple NOEL/NOAEL data has not
                     been developed or accepted by the
                     scientific community.

                     Issue 19
                      Comment summary: Is there a
                    reasonable way to use dose/response
                     data to derive criteria?
                      Response: Mathematical models for
                     deriving non-cancer based criteria are
                    available. However, they have not
                    gained wide acceptance in human risk
                    assessment Until various models have
                    been reviewed in greater detail, the
                    Agency uses the current approach,
                    based on that recommended by the
                    National Academy of Sciences, as the
                    most appropriate,
                    Issue 20
                      Comment summary: Confidence
                    Intervals or a range should be used in
                    deriving criteria,
                      Response; A workable method for
                    using confidence intervals in deriving
                    non-cancer based criteria has not been
                    developed. Given the many
                    uncertainties involved in this process,
                     the use of confidence intervals could be
                    misleading in simply considering
                    problems in statistical variation without
                    considering problems in species to
                     species conversion. Safety factors are
                    an accepted procedure and are used to
                    consider both problems in statistical
                    variability as well as problems in
                    species to species conversions and
                    individual susceptibility.
                    H. Exposure

                    Issue 21
                      Comment summary: Should non-
                    cancer criteria be based on all sources
                    of exposure because they are derived
                    from estimates of ADIs (acceptable
                    daily intake) which define total daily
                    acceptable doses for man?
                      Response: The methodology has been
                    revised so that estimates of total
exposure can be considered in deriving
criteria. Estimates of water and fish
consumption arc used to derive the
criteria. However, the criteria levels can
be modified by considering all routes of
exposure in the standard-setting
process. This approach may be
particularly desirable because exposure
conditions will probably vary markedly
on a regional basis.

Issue 22
  Comment Summary: If sufficient data
are not available on all sources of
exposure, can any reasonable
assumptions be made to factor in all
sources of exposure or can/should an
additional "uncertainty" factor be used?
  Response: When no reasonable
estimate can be made of contributions
from non-fish diet and from air, it can be
assumed that one-half of the exposure
comes from water and fish and one-half
comes from other sources. This is
equivalent to using an additional safety
factor of 2. It is recognized that the
inability to quantify all sources of
exposure adds an additional element of
uncertainty to the criteria.

/, General Issues
Issue 23                          .
  Comment Summary: With the
exception of recommending "good
scientific judgment," can specific
guidelines be given for accepting or
rejecting a study or set of studies as a
data base for criteria derivation?
  Response: Specific guidelines cannot
be given for accepting or refecting
studies. Scientific judgment must be
exercised in view of the magnitude of
the total evidence on the chemical or
chemicals under consideration. Chronic
data and appropriate exposure routes
are most desirable.
Issue 24
  Comment Summary: Is there a need to
individualize the criteria derivation
process so that the "nature of the toxic
agent and its mechanism of action" can
be more explicitly considered? If so,
how can this be accomplished?
  Response: The criteria derivation
process does consider as specifically as
possible the nature of the toxic agent
and, when known, the mechanism of
action.

Issue 25
  Comment Summary: Is the Stokinger-
Woodward model adequate for
converting inhalation dose data to
"equivalent oral doses," or should a
more sophisticated approach be used?
  Response: The derivation of water
quality criteria from inhalation data is
an admittedly tenuous process. The

-------
                  Federal Register / Vol. 45. No.  231 / Friday. November 28.  1980 / Notices
                                                                     79373
following guidelines have been added to
the methodology.
  Estimating equivalencies of dose/
response relationships from one route of
exposure to another introduces an
additional uncertainty in the derivation
of criteria. Consequently, whenever
possible, ambient water quality criteria
should be based on data involving oral
exposures. Even with oral data.
differences in dosing schedules and
vehicles can be problematic. If oral data
are insufficient, data from other routes
of exposure may be used in deriving
water quality criteria. \
  Inhalation data, including TLVs or
similar values, are the most common
alternative to oral data. Estimates of
equivalent doses can be made on the
basis of extensive pharmacokinetic data
for oral and inhalation routes, on the
basis of measurements of absorption
efficiency from ingested or inhaled
chemical, or on the basis of comparative
excretion data when the metabolic
pathways can be established to be
equivalent after oral or inhalation
dosing. When sufficient
pharmacokinetic data are available, the
use of accepted pharmacokinetic models
provides the most satisfactory approach
for dose conversions. However, if the
pharmacokinetic data are marginal or of
questionable quality, pharmacokinetic
modeling is inappropriate and may
result in an artifical sense of exactitude.
  The Stoldnger and Woodward (1958)
approach, or similar models which are
based on assumptions of breathing rate
and absorption efficiency, can be used
as alternatives when data are not
sufficient to Justify pharmacokinetic
principles. Consequently, in using the
Stokinger and Woodward or related
models, the uncertainties inherent in
each of the assumptions and the basis of
each assumption should be clearly
stated in the derivation of the criteria.
  The use of data involving other routes
of exposure to derive water quality
criteria should not be ruled out
However, as with Inhalation  data, an
attempt should be made to use accepted
toxicologic and pharmacokinetic
principles to estimate equivalent oral
doses. If simplifying assumptions are
used, their bases and limitations must
be clearly specified.
  Because of the uncertainties involved
in extrapolating from one route of
exposure to another and the consequent
limitations that this may place on the
derived criteria, the decision to disallow
such extrapolation and recommend no
criterion is highly judgmental and must
be made on a case-by-case basis. Such a
decision should balance the quantity
and quality of the available data against
a perceived risk to the human
population if no criteria is derived.

Issue 26

  Comment Summary: Can/should
criteria be qualitatively  or quantitatively
ranked in terms of their scientific
strength of validity? How could such a
ranking system be developed?
  Response: The Agency is presently
assessing the quality of the data base
supporting individual criteria. This will
eventually result in the development of
a ranking system of all the priority
pollutants.

TV. Response to Public Comments on
Methodology to Derive  Water Quality
Criteria

  The Carcinogen Assessement Group
(CAG) and the Environmental Criteria
and Assessment Office-Cincinnati
(ECAO-Cln.) cf the U.S. Environmental
Protection Agency (EPA) has reviewed
in detail the public comments on EPA's
methodology to derive water quality
criteria for carcinogens. Since the
majority of the comments are concerned
with the low-dose extrapolation
procedure and since they are closely
related to each other, an appendix is
presented which summarizes our new
procedure to derive water quality
criteria and the rationale for selecting
the procedure and compares the new
with the old procedure. Much of the
criticism has been directed toward
utilization of the one-hit linear model for
estimation of the risk. After
considerable input by a  peer review of
outside scientists, the multistage model
developed by Kenneth Crump has been
adopted in place of the one-hit model
extrapolation. The Appendix describes
the new multistage hit model Further
responses to the individual comments
are being presented below.

A. The One-Hit Model

Issue 1

  Comment sununa/y.-.Several
comments criticize  the one-hit model as
arbitrary, inappropriate, simplistic,
unrealistic. Inaccurate, not universally
accepted, and/or overly conservative.
  Response: The Agency has adopted a
new procedure for deriving water
quality criteria which is conceptually
similar to, but operationally more
systematic than the one-hit procedure
used previously by the Agency.
Although the criteria calculated by the
new procedure are  not appreciably
different than those calculated by the
old procedure as demonstrated in the
appendix, most of the general criticisms
do not apply to the new procedure.
Issue 2
  Comment summary: Comments
pointed out that the EPA has declined to
use the one-hit model under the federal
pesticide laws for heptachlor and
chlordane.
  Response: The commentor is correct
that the one-hit model was not used in
the chlordane-heptachlor suspension
hearings in 1375. However, in the
cancellation hearings, which were held
after the formation of the Carcinogen
Assessment Group and the adoption by
the Agency of the Interim Cancer
Assessment Guidelines and in the
proposed water quality criteria, one-hit
extrapolation model was used for risk
estimation. In the current final water
quality documents the "linearized"
multistage model is used; the
comparison between these two
approaches in the appendix to those
comments shows that the chlordane and
heptachlor data have the largest upward
curvature in the dose-response curve of
all the carcinogens in the water quality
list. For this reason the new approach
reduces the risk for chlordane and
heptachlor more than for the other
compounds. This example shows how
the new extrapolation procedure
compensates for the "overly
conservative" results of the one-hit
approach in cases where the dose-
response data is sharply concave
upward at low doses.

Issue 3
  Comment summary: The EPA's choice
of this model because ". .  . it gives
greater risk estimates than other
plausible models" (page 15978 of March
15, Federal Register] was criticized as
being a policy/political/social statement
rather than a scientific defense.
  Response: See the appendix for
reasons for selecting linear, non-
threshold models.

Issue 4
  Comment summary: The statement
that this model was endorsed by IRLG
(1979) was felt to have limited meaning
because this document has not yet been
reviewed and because the document is
merely a reiteration of policy.
  Response: The model was not selected
on the endorsement of IRLG. See
appendix.                        •

Issue 5
  Comment summary: In the
methodology (page 15978,  column 1, first
full paragraph of the March 15. Federal
Register], this model is scientifically
defended as being consistent with three
basic concepts in chemical
carcinogenesis:

-------
 7937-1
Federal  Register / Vol. 45. No. 231 / Friday, November 28, 1980  /  Notices
   a. The Linearity of the Doae-reaponae
 Curve for Mutagens—This Is challenged
 on the following points:
   Hie shape of the dose response curve
 In the low dose region cannot be
 determined.
   Not all assay systems give linear
 dose-response patterns.
   Some Ames tests are linear because
 the liver microsomes are added in a
 fixed amount and thus ". . . the laws of
 first order kinetics require • linear
 response to the variation in
 concentration* of die test substance as
 it la mediated by the activator,"
   b. Chemicals which are Mutagens are
 Likely to Induce Cancer—This is
 challenged on the basis that not all
 mutagens cause cancer. -
   c. Epidemiology Studies on Radiation,
 Cigarettes, and Aflatoxin show a Linear
 Dose-response Pattern—This is
 challenged on the following points:
   Radiation carclnogeniclty cannot be
 applied to chemical carcinogenicity
 because they act by different
 mechanisms.
   Not aU radiation dose-response data
 li linear.
   Smoking data are compounded by
 difficulties with cocarcinogens and other
 exposures,
   Aflatoxin data rely purely on
 estimated exposures.
   Response: [a] The commentor points
 out that even in mutagenesis test
 systems there is a level of mutagenic
 response that is too small to be detected
 and that below this level the shape of
 the dose-response curve cannot be
 measured. While this is true, the
Agency's point is that the mutagenesis
data available are fundamentally
 consistent with a linear no-threshold
mechanism of action. Another
 commentor has misinterpreted the
mutagenesis dose-response data. As
presented by the original authors, the
data show some residual mutagenic
activity at zero dose. This is interpreted
erroneously as being a threshold below
which no response occurs. Another
commentor supports the Agency's
contention that the mutagenesis dose-
response relationship is linear by giving
a possible explanation for the linearity.
  (b) The fact that chemicals which are
mutagenic are "likely" to Induce cancers
does not imply that "all" mutagens
cause cancer. Furthermore, those
mutagens which were not shown
experimentally to be carcinogenic could
not be accepted unequivocally as non-
carcinogenic because of the uncertainty
in the study outcome.
  (c) Both chemicals and radiation
cause DNA damage and subsequent
interference with the normal functioning
 of DNA, although the mechanisms for
                     causing this damage are different for
                     radiation and chemicals.

                     Issue 0
                       Comment summary: Several
                     comments stated that the possibility of
                     thresholds for at least some chemical
                     carcinogens is not unreasonable, should
                     be addressed in greater detail and/or
                     cannot be resolved at this time. The
                     possibility of assuming a threshold was
                     recommended for the following
                     compounds; chloroform, PCBs,
                     acryionitrile, hexachlorocyclohexane,
                     chlorinated benzenes, and chlorinated
                     •thanes.
                       Response: Currently there is no
                     satisfactory method for estimating the
                     low-dose carcinogenic risk to
                     "epigenetic" chemicals. Until the
                     mechanisms for such action are
                     understood on a case-by-caso basis to
                     the pout of being able to justify a
                     specific extrapolation procedure, the
                     linear, no-threshold concept will be
                     assumed to be valid. The "linearized"
                     multistage approach now used result in
                     lower risks than the older  "one-hit"
                     approach for compounds having a sharp
                     upward curvature.
                       For the specific chemicals referred to
                     in Issue 0, no evidence was presented in
                     support of a carcinogenic threshold dose
                     except for chloroform. Commentors
                     state that chloroform induces an
                     increased rate of cell proliferation,
                     which they implicitly equate with
                     carcinogenesis. at high doses because of
                     a cytotoxic response which la unrelated
                     to direct DNA Interaction and which
                     therefore is not expected to occur at low
                     doses. Three pieces of evidence are
                     cited in support of that position: (a)
                     chloroform Is not mutagenic in the Ames
                     tests: (b) at doses below IS mg/kg/day.
                     mice show no excess rate  of DNA
                     synthesis in kidney and liver tissue. This
                     excess is expected for a cytotoxic
                     response leading to cell proliferation; (c)
                     Roe et al (1979) on the basis of
                     responses in four strains of mice, has
                     established a no-carcinogenic effect
                     level of 17 mg/kg/day, whereas the
                     positive NCI experiment used by EPA
                     for the water criterion was carried out at
                     200-400 mg/kg/day.
                       Before the existence of a threshold for
                     chloroform can be established several
                     issues need to be resolved: (a) are the
                     no-effect levels in the DNA synthesis
                     studies and in Roe's observations real
                     phenomena or only artifacts occurring
                     simply because the limit of detection in
                     these studies was being reached? (b)
                     The relation between the cellular
                     proliferation, which is alleged to be
                     manifested by increased DNA synthesis,
                     and carcinogenesis is unclear, since in
                     the mouse strains used by NCI kidney
tumors do not occur and liver tumors do
whereas in the experiments cited by a
commentor both liver and kidney exhibit
DNA synthesis.

Issue 7

  Comment summary: A distinction
should be made between genetoxic and
epigenetic carcinogens based on
mutagenlcity data. These comments
imply that a threshold model would be
more appropriate for epigenetic
carcinogens,
  Response. While it la true that moat
carcinogens do Interact with DNA, there
are some compounds, such as phorbol
esters in mouse skin studies and
phenobarbital in rat liver, which are
incomplete carcinogens by themselves;
but require another substance to initiate
or promote their action. In these studies
the effects are unrelated to DNA
interactions and apparently involve
important recovery processes. This
newly-developing field is not yet well
enough understood to justify the use of a
particular dose-response extrapolation
model.

Issue 8

  Comment summary: Another group of
comments vigorously opposed the non-
threshold assumption used in the one-hit
model. Criticism of the non-threshold
assumption were most extensively
articulated by commenton which
contended that the non-threshold
assumption is:
  Contrary to experience and logic, to
what is known of biological systems,
and to existing scientific data and is a
product of the desire to obtain a simple
and easy-to-use method for criteria
derivation.
  A related comment contended that
thresholds are apparent for mutagens
and therefore—given the presumed
relationship of carcinogenicity to
mutageniclty—thresholds  should be
postulated for carcinogens.
 • Response: Commentors state that the
linear non-threshold model is: (a)
contrary to experience and logic; (b)
contrary to what la known about
biological systems; (c) contrary to
existing scientific data and (d) an
approach based on faith that could not
be disproved by any facts.
  (a) The linear non-threshold model
does not imply, as suggested by a
commentor, either that (a) cancer is
inevitable In the general public or in
heavily exposed industrial workers or
that (b) all substances are carcinogenic.
ll simply states that the probability of a
person getting cancer is proportional to
the amount of carcinogen  to which he is
exposed.

-------
                 Federal Register  /  Vol. 45. No. 231 / Friday, November 28. 1980 / Notices^
                                                                      79375
  : ' First order reaction processes are
   iion In biological systems especially
 ,-riutagenesls.
 (c) Dr. Blngham's article did not
 ivoeate a elgmoid dose-response curve
 preference to a linear curve, as stated
 . the-commentor. She stated that
 veral environmental factors can alter
 e dose-response relationship, and
 ould thereby change the curve
 hichever way it was described. In fact,
 T main point was that "until we
 iderstand more about the primary
 ircinogenle insult and its progression,
 edicling or estimating thresholds is
 iky," The Agency agrees with this
 inclusion.
 [d] The Agency agrees that it would
 extremely difficult to use negative
 idemiology data as proof that a
 rcinogenic threshold exists for a
 mpound having positive animal
 suits.         ,

 iue 9
 Comment summary:, Several
 mments critlzed the one-hit model
 cause it does not fit some
 perimental data as well as other
 idels. This was illustrated for
 ptachlor, chlordane, and aflatoxin and
 lorinated ethanes.
 Response: The new extrapolation
 slhod overcomes the difficulty In
 ting the model to the data because the
 iltistage model has enough flexibility
 fit any montonically increasing dose-
 jponse relationship. See also the
 iponse to Issue 16.

 uelO                          •
 Comment summary: The application
 the model was also criticized because
 lisregards data at all but one dose
 ei and fails to consider the results of
 ler experiments.
 Response: The new procedure does
 t have these shortcomings. See
 oendix.

 ue 11
 Comment summary: The highest
 ency factor to the exclusion of all
 er data should not be used in
 icrating criteria because this process
 :s not involve maximum-likely risk
 imates.                •      . •-.
 lesponse: In judging which of several
 mal studies to use  as the basis for the
 mtitative risk estimate, the quality of
 h study is considered as well as the
, nerical slope factor. As explained in
 preamble, an experiment with a
 all number of animals is rejected in
 or of a  larger experiment if the two
 •e a similar dose-response
 itionship. A similar rejection is also
 -de if an experiment is judged to be
 eliable for other reasons. Because of
 the strain, species, and sex differences,
 it is considered improper to calculate an
 average response across all animal
 species and designate this average as
 the carcinogenic potency for animals in
 general.
 Issue 12
  :Camment summary;". .., no"
 experiment, however large and well run,
 could ever reduce these estimates
 {criteria)."
  Response; In judging which of several
 animal studies to use as the basis for the
 quantitative risk estimate, the quality of
 each study is considered as well as the
 numerical slope factor. As explained in
 the preamble, an experiment with a
 small,number of animals ia rejected in
 favor of a larger experiment if the two
 have a similar dose-response
 relationship. A similar rejection is also
 made if an experiment is judged to be
 unreliable for other reasons. Because of
 the strain, species, and sex differences it
 is considered improper to calculate an
 average response across all animal
 species and  designate this average as
 the carcinogenicity potency for animals
 in general.
 Issue 13
  Comment summary: The EPA method
 is insensitive to reproducibility of the
 results, results at lower doses, and the
 number of animals per dose group.
  Response: In judging which of several
 animal studies to use as the basis for the
 quantitative risk estimate, the quality of
 each study is considered as well as the
 numerical slope factor. As explained in
 the preamble, an experiment with a
 small number of animals is rejected in
 favor.of a larger experiment if the two
 have a similar dose-response
 relationship, A similar rejection is also
made if an experiment is judged to be
unreliable for other reasons. Because of
the strain, species, and sex differences,
 it is considered improper to calculate an
average response across all animal
species and designate this average as
the carcinogenic potency for animals in
general.
Issue 14
  Comment summary: Several examples
are given of the failure of the one-hit
 model to predict cancer rates in humans
based on epidemlologic studies:
  Analyses of data on; chloroform,
carbon tetra chloride,
 tetracholoroethylene, aflatoxin,
chlordane, arsenic, and beryllium.
  In a summary of analyses of DDT,
dieldrin, and aflatoxin, it is indicated
that the one-hit model predicts an
incidence of 153,000 liver cancers per
year but that the observed response rate
 from all chemicals in only 3,000 to 4,000
 per year. A similar analysis is made of
 pollution exposure-cancer rates in the
 Sacramento River area.
   Response.' For chloroform, carbon.
 tetrachloride, and tetrachloroethylene
 the analysis assumed that all of the...
 workers were exposed at the TLV levels
 for their entire lifetime. In reality most
 workers are not exposed continuously to
 levels as high as the TLV and most work
 for only a few years at these jobs. This
 procedure overestimates the average
 lifetime exposure by at least a factor of
 10 and the risk estimates for the workers
 are too high because of exposure
 assumptions used by the commentor
 rather than solely because of an
 overestimated slope factor.
   For aflatoxin the commentor showed
 that the multistage model fits the
 observed human data more closely than
 the one-hit model. Therefore, that
 analysis partially justifies the revised
 procedure, although this compound is
 not on the water quality list,
   The criterion for arsenic was based on
 human data, which was linear with
1 dose. However,  none of the negative
 epidemiology studies in areas with high
 drinking water levels of arsenic was
 inconsistent with the model developed
 on the basis of the Taiwan skin cancer
 data.
   Commentors estimated that the
 annual number of cancer cases caused
 by beryllium intake is about 14,000.
 They gave no reason why this number is
 considered excessive considering that
 400,000 cases per year are observed
 from all causes.
 Issue IS
   Comment summary: Based on the
 above types of analyses, several
 comments recommended that
 epidemlologic data be used to test and/
 or modify risk estimates.
   Response: The Agency agrees that
 good epidemiological data should be
 used to estimate or modify risk
 estimates. The Agency always preferred
 using epidemiological data to the animal
 data in deriving  water quality criteria.
 Issue 16
   Comment summary: Some comments
 suggest that selection of a particular
 model should be left open and subject to
 the nature of the experimental data end
 epidemiologic or metabolic information.
   Response: The Agency does not agree
 that the selection of a particular model
 should be left open and subject to the
 nature of the experimental data for the
 following reasons. When behavior of the
 dose-response curve at low doses is not
 sufficiently understood, it is more
 appropriate to predetermine the low-

-------
 79376
Federal  Register / Vol. 45.  No. 231 / Friday.  November 28.  1980 / Notices
 dose extrapolation model. Considering
 the fact that all the mathematically
 analytic functions such as most of the
 parametric dose-response curves could
 be approximated by polynomials which
 are dominated by the higher order terms
 in the high-dose range, whereas they are
 vanishingly small in the low-dose
 regions it is not surprising to see that
 different dose-response models could fit
 well a set of high dose data while their
 low-dose extrapolations differ
 drastically. Therefore, the selection of
 the extrapolation model should be based
 on knowledge of carcinogenic
 mechanisms (even though limited and
 debatable) rather than being determined
 solely by the high dose behavior of the
 dose-response curve.
 Issue 17
  Comment summary: Other comments
 suggested that the one-hit model should
 be used only in  the absence of data
 suggesting that other models give a
 better flt.
  Response: See response to Issue 16
 and the appendix.
 Issue 18
  Comment summary: Two comments
 recommend that several models used for
 analysis along with appropriate
 confidence intervals would more
 objectively reflect the state of scientific
 knowledge.
  Response: The inclusion of several
 arbitrary models in order to get a range
 of risk estimates would add no
 additional scientific information while
 at the same time would create confusion
 and thereby undermine the utility of risk
 estimates. The model chosen by the
Agency is regarded as giving a plausible
 upper limit to the risk.

 Issue li
  Comment summary: A general
 discussion of alternative models is given
 in some of the comments. Specific
 models recommended include: Logistic;
Probit; Multi-hit; Mantel and Bryan;
 Weibull; and Pharmacokinetic.
  Response: The Inclusion of several
 arbitary models In order to get a range
 of risk estimates would add no
 additional scientific information while
 at the same time would create confusion
 and thereby undermine the utility of risk
 estimates. The model chosen by the
Agency is regarded as giving a plausible
 upper limit to the risk.

B. Use of Confidence Intervals
 Issue 20

  Comment summary: Confidence
 intervals or a range should be used in
 deriving criteria.
                       Response: The Agency feels that the
                     statistical confidence intervals should
                     not be used to express the range of
                     uncertainty of the criteria because this
                     range does not include major
                     uncertainties which are not quantifiable,
                     such as species differences in
                     metabolism, diet, target organ
                     specificity, and other biological
                     variables.

                     C. Species Conversion Factor (Wr
                     Issue 21

                       Comment summary: Comments
                     suggested that this factor may not be
                     appropriate for carcinogens because; (a)
                     DNA repair rates appear to be inversely
                     proportional to body weight, and (b)
                     mixed-function oxidase activity, which
                     may activate carcinogens, is higher in
                     rodents than in man. Examples were
                     given indicating that man is less
                     sensitive than experimental mammals to
                     chloroform, aflatoxln, and vinyl
                     chloride.
                       Response: Although some commentors
                     discussed reasons why the species
                     conversion factor, (70/W)V&, may not be
                     appropriate for particular compounds,
                     no suggestion was made for an
                     alternative method which would be
                     valid in general Commentors suggested
                     that mixed-function oxidase activity Is
                     lower in humans than in rodents and
                     that humans metabolize chloroform less
                     completely than animals, but facts like
                     these, even if quantified would have
                     uncertain implications to carcinogenic
                     potency in general because increased
                     metabolic activity could both enhance
                     carcinogenic potency by  "inactivating"
                     the agent.
                       The fraction of a compound  (e.g,
                     chloroform) unmetabolized may have no
                     relation at all to the amount of active
                     metablite formed. In the general method,
                     the cube root factor is Intended to
                     account only for the body size difference
                     between animal speciea as it relates to
                     the availability of the chemical to the
                     body tissues. Any specific knowledge
                     available on metabolism differences
                     would have to be incorporated as an
                     additional factor if it could be  directly
                     related to cancer incidence. In general,
                     mixed-function oxidase activity has no
                     clear relation to cancer occurrence,
                     therefore, cannot be included in the
                     general approach.
                       The best apporach for  checking the
                     validity of the species conversion factor
                     is to correlate carcinogenic potency of
                     agents in animals with that in  humans
                     where suitable information is available.
                     This was done in a preliminary fashion
                     by Messelson [quoted by one
commentor) and is currently being
investigated by the Agency,
Issue 22                    .

  Comment summary: Data on
comparative metabolism should be used.
whenever possible, to modify the risk
estimate.
  Response: See Issue 21 for response to
chloroform metabolism issue. The
Agency acknowledges that species
differences in metabolism should be
considered in all cases where the data
can be interpreted as being relevant to  '
carcinogenicity. In the methodology
description the appropriate place to
incorporate this information is in the
factor r, called the absorption fraction.
D, Time-To-TumorData

Issue 23

  Comment summary: Some comments
stated that the EPA's modification of the
one-hit model does not consider the
time-to-tumor concept
  Response: The time-to-rumor concept
was incorporated  in the one-hit
procedure by using the model P=l—exp
{—bd t*J where t is the average fraction
of a lifetime the tumor was observed
and is-a] so incorported into the current
approach. If sufficiently well defined
time-to-tumor data are available, a more
refined model would be used.

Issue 24
  Comment summary: Other comments
contended that, because of the
relationship between dose and latency,
even potential carcinogens will not
induce tumors in a normal lifespan.
Examples were, given for beryllium and
arsenic.
  Response: The arguments given in this
comment do not invalidate the criteria
which are associated with a lifetime risk
of 10~*. The arguments given in the
comment proceed as follows:
  Let F[d,t) be the probability of cancer
by age t at exposure d. F is a monotonic
increasing function of both variables t
and d. Let d, be the exposure associated
with the lifetime risk of cancer 10~*
obtained by solving for d from the
equation F(d,t) • io~f and t = 70 which
is taken as the average lifespan. Based
on the arsenic risk assessment by the
CAG the comment argues that at
exposure d«= 0.002 fig/liter (where dc is
associated with a  lifetime risk of 1CT6),
the median age would be 2,636 years
before cancer can occur, where 2,636 is
obtained by solving for I from the
equation F(d0,t) =  0,5.
  Therefore, if the water concentration
is 0.002 tig/I the risk at 70 years is 10~*
and at 2,636 years it would be 0.5. These

-------
                  Federal  Register /  Vol. -45. No. 231 / Friday. November 28. 1980 / Notices
                                                                      79377
are ~ it inconsistent statements, as
inv   J by the commentor.
      nent summary: One comment
sug,  ^ted that studies in which t is less
than 0.75 may not be useful because of
insufficient time for tumor development.
  Response: The t3 factor is necessary
when the carcinogenic response is so
strc -g that the animals die prematurely
of tunion. Hits Is true regardless of
whether the median time of death from
tumors Is greater or less than three-
fourths of their natural llfespan.

Issue 26
  Comment summary: Another comment
raised questions about the experimental
difficulties of Druckery's work in
precisely determining time-to-tumor
development and the need to correct
tlme-to-tumpr data for the degree of
maUgnacy of the tumor.
  Response:The tlme-to-tumor data is
used only when there was an early
terminal sacrifice. In this case the full
spectrum of tumor development is
observable histologically and the
difficulty of observing the precise time
of tumor development is not
encountered.
E. Mutagenicity Data
Issue 27
  Comment summary: Mutagenicity
data should be given greater weight to
determine potential cardnogenidty
especially when mammalian bioassays
or epidemiology data are lacking.
  R(  -wise: See response to Issue 28.
Issue 28
  Comment summary: The commentor
describes a method for using results
from short-term tests, such as the Ames
test or the hamster embryo in vitro
transformation test, to perform       v
quantitative cardnogenidty risk
assessment
  Response: The Agency does not
regard results from short term
mutagenidty tests, even those from a
test battery using several organisms,
equivalent to chronic whole animal
bioassays for cardnogenidty because of
the inherent differences between the
test systems utilized (i.e., bacteria and
cell cultures versus whole animals, in
which all the metabolic, distribution and
excretion systems of the body are
intact). Until correlations between the
carcinogenicity and mutagenidty of
agents become better understood and
more widely accepted, and until the
data base is more extensive,  the Agency
is not justified in making quantitative
assessments of carcinogenicity based
solely on mutagenidty test results  and
structure activity relationships. In     *
decisions regarding the carcinogenicity
of agents, the Agency currently uses
short term bioassay tests only to support
equivocal findings of long term animal
bioassays and human studies.

F. Epidemiology Data
Issue 29
  Comment summary. Epidemiology
data on societies other than the U.S.A.
should not be used because of
"dissimilar and possibly controlling
variables."
  Response: The CAG feels that
epidemiological data on societies other
than the U.S.A. population can be used
as long as care Is taken in interpreting
and using the data.

Issue 30
  Comment summary: Citing criteria for
arsenic and cadmium, the comment
states that: "Valid epidemiological
studies exploring a cause-and-'effect
relationship between exposure to a
substance and disease must avoid a
number of flaws: bias, confounding
factors, and the confusion of chance
associations with casual relationships.
The epidemiological studies used by the
Agency in criterion formulation fail to
avoid these flaws."
  Response: While nearly every
epidemiologic study contains flaws in '
the sdentiflc sense, a regulatory agency
must interpret all data available, making
judgments as to whether the studies
were too flawed for proper conclusions.
In determining the carcinogenicity of a
substance, the Agency is sensitive to the
need to find human populations who
have been exposed to other agents also.
The appearance of rare types of cancers
and/or a dose-response trend, however.
often provlde(s) very positive evidence
of carcinogenidty. Such is  the case with
the Taiwan drinking water survey. Here.
where artesian well water with a high
concentration of arsenic has been used
for more  than 60 years, a high
correlation between amount of arsenic
and skin  cancer was  found In addition,
the pre-cancerous skin conditions were
pathonbmic of arsenic exposure, so that
there was little chance that the cancers
were caused by another agent.
Furthermore, the skin cancer is of a rare
form that was virtually unknown in
parts of Taiwan where the drinking
water arsenic content was small. In
addition, a positive association between
arsenic level in drinking water and the
prevalence of skin cancer has been
reported  in at least three other areas in
the world.
  Cadmium is an unusual situation in
that five  independent populations
showed an excess of prostate cancer.
Even though each study is inconclusive
by itself for the reasons cited, a chance
occurrence of this finding is exceedingly
unlikely.
  While the effect of many possible
confounding factors, especially
concomitant exposure to unknown
chemicals, cannot be accurately
determined, the Agency has the
responsibility of estimating criteria
levels with the best information
available.
C. Qualitative Determination of
Carcinogenicity

Issue 31
  Comment summary: The commentor
states that: "A substance is currently
considered to be  carcinogenic if it
produces a statistically significantly
higher than normal incidence of tumors
in treated animals in a single test. Such
a result is inconclusive, because of the
problems of false positives.
  Response: In establishing a false
negative rate of P<0.05 the commentors
correctly point out that the false positive
rate is rather high. However, the careful
review of other information about the
compound reduces the effective false
positive rate.
Issue 32
  Comment summary: The decision to
label a compound "a suspect human
carcinogen and therefore a potential
human carcinogen" based on
tumorigenicity in experimental
mammals has not been validated.
  Several comments from the initial
publication of the methodology made a
similar critidsm.
  Response: Among public health
authorities it is widely accepted that the
positive results in chronic animal
bioassays indicate that the agent poses
a potential risk for human
carcinogenicity. This attitude is
thoroughly summarized in the IRLG
report (Jour. Natl. Cancer Inst 63:241,
1979). In addition, a  review by Tomatis
(Am. Rev. Pharmacol. Toxlcol. 79: 511,
1979) emphasized the value of rodent
bioassays in predicting human
carcinogenic risk.
  Since 1976,  (41FR 21402) the EPA has
been following the same regulatory
philosophy in evaluating carcinogenic
hazards. Therefore,  contrary to the
comments, the EPA has not been acting
unilaterally without adequate public
notice.

Issue  33
  Comment summary: The commentor
quotes a WHO publication: "It would be
unwise to classify a substance as a

-------
 79370
                   Federal Register / Vol. 45.  No. 231 /  Friday.  November 28, 1980 / Notices
 carcinogen solely on the basis of a
 species or strain-specific increased
 incidence of tumors of a kind that occur
 spontaneously with high frequency."
   Response.- The CAG agrees partially
 with the comment However, even if the
 spontaneous Incidence is high, a
 statistically significant enhancement of
 the tumor incidence is of concern, and
 other evidence for the compound should
 be evaluated for consistency with that
 finding.
 Issue 34
   Comment summary: Citing PAH as an
 example, conunentors state that the
 documents have been inconsistent in
 qualitative determinations of
 carcinogen! city.
   Response: In cases such as PAH
 where one criterion has to be set for an
 entire class of compounds, the Agency
 does not state that each chemical in  the
 class is a carcinogen,  as Implied by the
 commentor. Therefore, the PAH
 example cited by the commentor does
 not show that the Agency is  Inconsistent
 in classifying compounds as
 carcinogenic.
  The intended interpretation of the
 criterion is that the risk is less than HT*
 whenever the total concentration of all
 PAH compounds in water is less then
 the criterion. In a hypothetical case
 where all of the PAH compounds In a
 sample are non-carcinogenic, the
 criterion would be too strict;  however,
 this situation seldom occurs.  In most
 cases where PAH is detected, a mixture
 of compounds occurs and hi calculating
 the criterion the assumption is made
 that all components have the same
 carcinogenic potency as benzdkt-
 pyrene.
 Isaue 35
  Comment summary: Bis(2-
 chloroisopropyljether  (BCIE)  yielded
 negative results in an NCI bioassay.
 Nonetheless, the cancer based criteria
 based on an upper limit of the true
 response rate was calculated because of
 the structural similarity of BCIE to other
 carcinogenic chloroalkyl ethers. The
 commentor states that this is
 inappropriate.
  Response: As a response to the public
 comments, the Agency has changed its
 interpretation of data on bis(2-
 ehlorolsopropyljether (BCIE). It Is no
longer considered to be carcinogenic
and, therefore, a criterion based on
 carcinogenic data is not calculated.
Issue 36
  Comment summary: Using  the criteria
 for chromium and asbestos as examples,
 the commentors state that data on
 inhalation carcinogenicity should not be
 used to derive criteria if oral
 carcinogenicity tests are negative.
   Response: The criteria for chromium
 (Cr) was derived on a marginally
 significant digestive cancer Incidence
 which occurred from inhalation
 exposure (Enterline epidemiology
 study). The digestive system cancer is
 assumed to be caused by chromium (Cr)
 which is removed from the respiratory
 system by mucociliary action and then
 swallowed. This is comparable to
 exposure to chromium in drinking water.
 Cr (VI) has not been adequately tested
 .for its carcinogenic potential in animals;
 therefore, further studies to assess the
 caicinogenidty of Cr (VIJ by the oral
 route would be desirable and necessary
 before it is concluded that oral tests are
 negative.
   Asbestos has been shown to cause
 peritonea] mesothelioma in humans and
 is also associated with a significant
 increase In human gastrointestinal
 cancers. These are caused by inhaled
 asbestos. Since up to 99 percent of the
 inhaled  asbestos is eventually
 swallowed, the Agency feels that
 asbestos-contaminated water could
 cause the same type of gastrointestinal
 cancers as inhaled asbestos.
 H. Joint Action/Cocarcinogen/city

 Issue 37

  'Comment summary: The commentors
 emphasized the potential importance of
 cocarcinogenlcity and possible
 synergisUc effects among carcinogens.
  Response: The potential importance of
 cocarcinogenicity and possible
 synergistic effects among carcinogens
 has not been addressed by the CAG hi
 deriving water quality criteria, since
 sufficient data is not available at this
 time to make decisive judgments related
 to these issues.

/. Site Specific vs.  Total Tumors
 Issue 38

  Comment summary: No public
 comments specifically addressed this   '
issue. However, the methodology
committee should discuss the
appropriateness of using data on total
 tumors for quantitative risk assessment
  Response: Since chemicals generally
exert their carcinogenic effects at
specific organ sites, the incidence of
 tumors at the responding sites  is the
most relevant information to consider in
making either qualitative or quantitative
evaluations of hazard. The instances
where the tumor incidence at all sites
combined is elevated, but no one site or
group of sites is significantly increased,
are regarded as weak evidence of
carcinogenicity.
/. General Issues
Issue 39

  Comment summary: Single unverified
bioassays should not be used for
establishing criteria for
dichlorobenzene.
  Response: The comment does not
refer to carcinogenicity data since no
carcinogenic information was available
on dichlorobenzene.

Issue 40

  Comment summary: Some comments
expressed concern with the types of
studies used to derive criteria. Another
comment implies that only data
published in referenced journals should
be used. Two commentors recommend
that explicit reasons be developed for
accepting or rejecting studies.
  Response: The evaluations of
bioassay studies for carcinogenicity by
the CAG is sufficiently detailed to be
equivalent to that given in peer
reviewed journals.
Issue 41

  Comment summary: The EPA "has
used animal studies without adequately
considering the nature of the toxic agent
and its mechanism of action, the
conditions of exposure, or the
physiological characteristics of the test
organism."
  Response: The Agency routinely
considers all of the available
lexicological data cited by the
commentor and agrees that these factors
are important.

Issue 42

  Comment summary: Several
comments questioned the
appropriateness of using studies from
one route of exposure—particularly
inhalation—to establish criteria for
ingestion.
  Response: If a given chemical induced
a carcinogenic effect by inhalation at a
distant site, it is likely that the
compound could also produce a
carcinogenic effect by other routes of
administration. Therefore, the Agency
considers it appropriate to use
inhalation data to derive criteria for
ingestion, recognizing the  difficulty of
determining the dose.
Issue 43

  Comment summary: Treating all of the
proposed criteria as if they were based
upon equally valid  data is not
scientifically sound: EPA must make
explicit the nature, extent, and quality of
the data utilized to estimate criteria.
  Response: The Agency has indicated
which criteria should be regarded as

-------
                      Federal Register /  Vol. 45. No. 231  / Friday. November 28.1980  /  Notices
                                                                              79379
 marginal based on the nature of the
 available data.
 Issue 44
   Comment summary: Criteria based on
 carcinogenic effects might not be
 adquate to protect humans from
 mutagenic, teratogenlc, or other toxic
 effects.
   Response:^be U.S. EPA Office of
 Health and Environmental Assessment
 Is currently developing guidelines for
 estimating the human risk to substances
 producing mutagenic. teratogenlc, and
 reproductive effects. \
Appendix

/. An Improved Procedure for Deriving Water
Quality Criteria

  As discussed In the methodology document
(1981)* the Carcinogen Assessment Group
(GAG) has adopted a new prodedure which is
more systematic than the one-hit procedure
used previously by the GAG for calculating
the water quality criteria. The model selected
for the low dose extrapolation is given by
P(d}=l-exp[-(q.+q,d+  . . . +q,dk]]
At low doses, the upper confidence limit for
the extra risk
                                       A(d)  - P(d)  -P(o)
                                                  1  -P(o)
 has  the  form
                                Au(d)  •  l-«xp
That is, the risk A«(d) Is always linerarly
related to d at low doses. The constant qi*
corresponding to the 95 percent upper
confidence limit for A(dj is taken as the
carcinogenic potency for calculating the
water quality criteria.
  Instead of extrapolating with the one-hit
model based on the lowest dose group
showing statistically slgniflcp.nl response as
previously used by the GAG. the new
procedure is employed because of the
following reasons: (1) the procedure Is more
systematic: (2) It invokes fewer arbitrary
assumptions; (3) the assumption of the low-
dose linearity Is not essential in the use of the
model and; (4) it incorporates data from all
of the dose groups which are consistent with
the multistage model At the same time, it is
conceptually consistent with the linear non-
threshold concept on which the one-hit
procedure was based.
  The Agency recognizes that there Is no
really solid scientific basis for any
mathematical extrapolation model which •
relates carcinogen exposure to cancer risks at
the extremely low level of concentration that
must be dealt with tn evaluating the
environmental hazards. For practical reasons
such low levels of risk cannot be measured
directly either by animal experiments or by
epidemlologlc studies. We must therefor*,
depend on our current understanding of the
mechanisms of carcinogenesis for guidance
as to which risk model to use. At the present'
time the domlnat view of the carcinogenic
process Involves the concept that most agents
that cause cancer also cause Irreversible
damage to DNA. This position Is reflected by
the fact that a very large proportion of agents
that cause cancer are also mutagenic.
  There is reason to expect the quanta!  type
of biological response that is characteristic of
mutagenesis is associated with a linear non-
   -.hold dose-response relationship. Indeed
   :-;• is substantial evidence from
   •: genesis studies with both ionizing
  ' Carcinogen Assessment Croup's Procedure for
Calculating Water Quality Criteria. Updated 1981.
radiation and a wide variety of chemicals
that this type of dose-response model is the
appropriate one to use. This Is particularly
true at the lower end of the dose-response
curve; al higher doses, there can be an
upward curvature probably reflecting non-
threshold dose-response relationships. The
linear non-threshold model is also consistent
with the relatively few epidemiologies!
studies of cancer responses to specific agents
that contain enough Information to make the
evaluation possible (e.g., radiation induced
leukemia, breast and thyroid cancer, and skin
cancer Induced by .arsenic In drinking water.
and liver cancer. Induced by aflatoxin in the
diet).  There is also some evidence from
animal experiments that is consistent with
the linear non-threshold model (e.g., liver
tumors induced In mice by 2-
acetylaminofluorene in the large scale
EDmStudy at the National Center of
Toxlcological Research and the initiation
stage  of the two-stage  carcinogenesis model
in the rat liver and the mouse skin].
  Because it has the best, albeit  limited,
scientific basis of any  of the current
mathematical extrapolation models, the
linear non-threshold model has been adopted
as the primary basis for risk extrapolation to
low levels of the dose-relationship. The risk
estimates made with this model should be
regarded as conservative, representing the
most plausible upper limit for risk. I.e.. the
true risk is not likely to be higher than the
estimate but it could be smaller.

//. Comparison of the new Procedure with the
Old (One-Hit) Procedure
  The Agency had previously calculated the
slope  b based on the one-hit model P=l-exp-
bd, using only the data from the lowest dose
group where the Incidence rate is statistically
significantly different from the control group
(see Federal Register. Part V. Thursday.
March 15.1979). The point estimate b was
taken as the carcinogenic potency for the
compound. Unlike the  new procedure, the
upper confidence limit was not used because
the GAG recognized that the one-hit model is
 usually conservative at low doses and thus
 the point estimate b of the slope was
 considered as an upper limit of the true
 carcinogenic potency. This ad hoc approach'
"was used because it Is simple and easy to
 understand.
   Since b was considered an upper limit In
 an ad hoc sense, it would be interesting to
 compare the new procedure with the one-hit
 procedure by calculating the ratio of two
 carcinogenic potencies b/q/ for 21 chemical
 compounds In the Proposed Water Quality
 Criteria Documents which have data from at
 least three dose groups. Except for chlordane
 and heptachlor the new procedure agrees
 with the one-hit procedure within a factor of
 2. When the one-hit procedure is modified
 (Table 1) the two procedures become
 comparable. Therefore, the old procedure
 could be used as a simple and quick way of
 estimating the carcinogenic potency.

   Table 1.—Ratio of Carcinogenic Potencies
                                                      Chemicals
                                                                          b/q,'
                                                                                        CMortan
 HeplacNor	..	
 Caibon tttfechlui no..
 PAH	
 HCB—
 1.2-OichlonMthane—
 Acfytoflit/ito •*._.•>.-.««-
 Henehkxoethane—
 2.4.e-TncMoroeh
-------