[Federal Register Volume 69, Number 85 (Monday, May 3, 2004)]
[Pages 24164-24168]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 04-9946]



Centers for Disease Control and Prevention

Interim Recommendations for Airborne Exposure Limits for Chemical 
Warfare Agents H and HD (Sulfur Mustard)

AGENCY: Centers for Disease Control and Prevention (CDC), Public Health 
Service, Department of Health and Human Services (HHS).

ACTION: Notice of interim recommendations for airborne exposure limits 
for chemical warfare agents H and HD (sulfur mustard).


SUMMARY: Agents H and HD are stored and are being destroyed by the 
Department of Defense (DoD). Public Law 99-145 (50 U.S.C. 1521) 
mandates that the Secretary of Defense carry out the destruction of the 
United States' stockpile of lethal chemical agents and munitions. 
Public Law 91-121 and Public Law 91-441 (50 U.S.C. 1512) mandate that, 
prior to the disposal of any such agent within the United States, the 
Secretary of Defense implement any precautionary measures recommended 
by the Secretary of the Department of Health and Human Services (HHS) 
to protect the public health. This notice provides CDC's interim 
recommendations for worker and general population airborne exposure 
limits (AELs) for sulfur mustard. These revised exposure limits replace 
CDC's previously recommended AELs originally issued in 1988. These 
limits are being issued as interim criteria pending improved 
characterization of carcinogenic potential associated with sulfur 

EFFECTIVE DATE: July 1, 2005. An implementation period is necessary to 
allow the DoD to make program adjustments and allow time for changes to 
environmental permits as required.

FOR FURTHER INFORMATION CONTACT: Dr. Paul Joe, Chief Medical Officer, 
Environmental Public Health Readiness Branch, Division of Emergency and 
Environmental Health Services, National Center for Environmental 
Health, CDC, 4770 Buford Highway, Mail Stop F-16, Atlanta, Georgia 

SUPPLEMENTARY INFORMATION: On July 22, 2003, CDC published 68 FR 43356, 
``Proposed Airborne Exposure Limits for Chemical Warfare Agents H, HD, 
HT (Sulfur Mustard)'' \1\ seeking public comment. Today's notice 
discusses major comments received, describes decisions regarding the 
public comments, and provides interim recommendations. CDC received 
comments from the U.S. Army, the State of Utah, the State of Colorado, 
and one employee union.

    \1\ The toxicity data for agent T are inadequate for setting 
exposure limits. The very low vapor pressure for agent T precludes 
it as a vapor under normal ambient conditions. For sulfur mustard 
and T mixtures, air monitoring for sulfur mustard alone should be 
sufficient under most circumstances to prevent airborne exposure to 

    The comments fell into the following general categories: risk 
management assumptions used in CDC's deliberations, selection of 
uncertainty factors, determination of the cancer potency factor for the 
mustard AELs, and practical concerns of conducting air monitoring at 
the lower exposure limits. The key comments potentially impacting CDC's 
recommendations are summarized and discussed below:
    1. One reviewer remarked that the 5-minute ceiling (Ceiling-5M) may 
require too short of an analytical cycle for use with dual-agent air 
monitoring instrumentation.
    Discussion: The Ceiling-5M was defined to provide a ceiling value 
for near-real-time (NRT) corrective action that would protect worker 
health in the short term and meet the long-term goal of keeping the 
carcinogenicity risk below one in one million. The 8-hour time-weighted 
average (TWA) exposure limit recommended by CDC in 1988 was implemented 
by the chemical demilitarization program as a ceiling value, monitored 
by NRT instruments having a sampling and analysis cycle time of under 5 
minutes. CDC's proposal sought to reflect this conservative 
implementation of the 1988 criteria.
    CDC closely examined the various implicit exposure doses, measured 
in terms of concentration multiplied by time of exposure (Ct), for 
various potential exposure scenarios. The ceiling-5M was based upon the 
analytic cycle times used in the stockpile demilitarization program. 
Longer sampling and analytic cycle times, such as those used in the 
monitoring programs for chemical agent storage facilities or 
nonstockpile program, could be considered in a similar manner, that is, 
by evaluating the effect on the Ct by changing duration of potential 
exposure with varying instrument cycle times.
    CDC examined the implication of applying the ceiling-5M agent 
concentration with cycle times greater than 5 minutes. Comments 
received from the Army, indicated that the dual agent monitors use 
cycle times of up to 10 minutes. Accordingly, CDC reviewed the impact 
of using 10- to 15-minute cycle times at the same concentration used 
with the ceiling-5M. Both the short-term and long-term health 
protection goals were met; that is, the effective dose or Ct associated 
at this level and duration are still well under the Ct for the acute 
threshold of effects level (referenced in the July 22, 2003, support 
document for the proposed sulfur mustard AELs) and the carcinogenicity 
risk per episode would be well under one in one million.
    The above analysis would suggest that a longer analytic cycle time, 
even up to the 15 minutes, associated with the Army's NRT monitoring 
definitions, would be acceptable at the

[[Page 24165]]

concentration proposed with the ceiling-5M. However, real-world leaks, 
spills, or other unplanned agent releases do not follow a defined 
pattern of gradual airborne concentration increase. The first cycle of 
a monitoring alarm could be at much higher concentrations than the 
ceiling-5M. Consequently, to limit potential agent exposure durations 
at higher level exposures, analytic cycle time should be kept as short 
as practicable.
    The final factor considered in CDC's review of this issue is the 
overall risk management implication of modifying the implied cycle time 
associated with the ceiling AEL. Clearly, the degree of protectiveness 
increases as the cycle time decreases, assuming all other quality 
control criteria remain constant. However, if programmatic delays or 
extraordinary new personnel protective measures are introduced as 
interim measures in the pursuit of more ideal monitoring capabilities, 
overall risk could increase to both workers and the public.
    In summary, CDC believes that the proposed ceiling-5M was overly 
proscriptive and possibly counterproductive. Accordingly, CDC 
redesignated this AEL as a 15-minute short-term exposure limit (STEL). 
The concentration value, 0.003 mg/m\3\, from the ceiling-5M is 
retained. This STEL is to be monitored with NRT technology using the 
shortest practicable instrument cycle time. For the maximum 15-minute 
duration of the STEL, the Ct is 0.045 mg-min/m\3\.
    2. One reviewer remarked that using the proposed general population 
limit (GPL) for worker protection could result in excessive false-
positive situations and attendant disruptions wherever significant 
interferences might be located.
    Discussion: The GPL is a criterion that is set to protect the 
general public. Community exposure limits are set lower than worker 
limits to reflect wider variation in human susceptibility than that of 
the healthy worker population.
    CDC premised its proposal to use the new GPL as a worker protection 
criterion on two basic considerations. First, because the GPL is 
designed to protect the community, it would also be adequate for a 
worker population. Second, CDC believed that historic monitoring for 
the GPL for demilitarization perimeter monitoring similarly could be 
implemented in worker locations to accommodate longer 12-hour shifts.
    As discussed in CDC's proposal, the GPL for sulfur mustard was 
driven largely by the goal of protecting the public at a cancer risk 
level of less than one cancer incidence in a million exposures at the 
GPL for 3 continuous years, a risk level that is considered to be 
negligible. Three years was chosen for the duration of the potential 
exposure at a GPL because it was believed to be the maximum duration of 
a campaign where sulfur mustard munitions would be handled and 
processed for destruction on a continuing basis. This assumed exposure 
scenario is conservative for both the public and workers for a number 
of reasons:
     No one worker works continuously for 3 years; 
actual time at work is probably well under one-third of all available 
hours per year when weekends, holidays, and vacations are considered.
     Demilitarization plant workers, storage site 
workers, non-stockpile site workers, or others who might reasonably be 
exposed to chemical agent do not remain stationary at one duty location 
for extended periods.
     Similarly, the individuals within the general 
community would not normally be anticipated to stay at one location 
continuously for 3 years.
     Varying meteorological conditions would preclude 
constant exposure conditions.
     With the rigorous active demilitarization site 
monitoring and the ongoing routine storage site inspection program, 
unplanned releases of chemical agent are unlikely to be sustained for 
any significant duration.
     CDC assumed exposure at the full GPL in its 
carcinogenicity evaluation even though detection at this level would 
result in investigation and remedial action. Typically, risk assessment 
professionals use some fraction of a ``practical quantification limit'' 
or detection level.
    The above mitigating factors suggest that long-term exposure 
scenarios (up to 3 years) used to estimate sulfur mustard 
carcinogenicity review overstates the true risk. Accordingly, CDC 
recommends retaining the proposed GPL for perimeter monitoring stations 
at demilitarization facilities and evaluation of the allowable stack 
    For worker protection against low-level exposure, CDC now 
recommends a separate 8-hour TWA for a worker protection limit (WPL) 
rather than applying the GPL as originally proposed. In the earlier 
proposal for mustard AELs, CDC investigated the development of a WPL 
using the Environmental Protection Agency (EPA) Categorical Regression 
(CatReg) method. The value derived from this method is 0.0003 mg/m\3\. 
This value is in reasonably close agreement with the U.S. Army Center 
for Health Promotion and Preventive Medicine (CHPPM) reference 
concentration-derived (RfC) WPL of 0.0004 mg/m\3\ and the Agency for 
Toxic Substances and Disease Registry (ATSDR) acute inhalation minimum 
risk level (MRL) \2\ of 0.0007 mg/m3 (1,2). CDC believes 
that the CHPPM-recommended value for an 8-hour TWA is protective for 
noncarcinogenic effects and should be implemented for worker 

    \2\ ATSDR defines an MRL as ``an estimate of daily human 
exposure to a substance that is likely to be without appreciable 
risk of adverse noncancer health effects over a specified route and 
duration of exposure.'' ATSDR also developed an intermediate MRL 
(continuous exposure for up to 1 year) for sulfur mustard at a value 
of 0.00002 mg/m\3\ that is numerically equivalent to the interim GPL 
recommended herein.

    3. The Army noted that, although CDC specified that the proposed 
AELs were developed for and based upon agent stockpile demilitarization 
practice, other non-stockpile and storage situations existed to which 
the AELs would be applied within other Army programs. Illustrations of 
a number of such situations and some suggested resolutions were 
provided for CDC's consideration.
    Discussion: In CDC's proposal, the use of Ct evaluations was 
emphasized as an indication of potential acute exposure dose. For 
potential applications beyond strict stockpile demilitarization, 
adjustments to implementation of AELs might be warranted on the bases 
of site-specific or activity-specific conditions. However, any such 
potential AEL implementation and adjustment for site-specific 
conditions must ensure that the new monitoring action level protects at 
the potential exposure dose (Ct) so that the recommended 8-hour WPL is 
not exceeded. Also, any NRT monitors should not have action levels set 
above the recommended STEL.
    4. Two reviewers commented that CDC's selection of the National 
Academy of Science (NAS) cancer potency factor (CPF) was inappropriate 
because the benzo-a-pyrene (BaP) index value used was based upon oral, 
not inhalation, exposure. They also believed that CDC should use the 
30-year exposure assumption described in EPA's risk assessment 
    Discussion: To estimate cancer risk, exposure assumptions and a 
numeric estimate of the potency of carcinogenicity of a substance are 
necessary. The reviewers believed that CDC should have used a 30-year 
duration for such exposure at the lifetime adjusted daily dose. CDC 
appreciates the general desirability to be

[[Page 24166]]

consistent with established guidelines in risk assessment, but EPA has 
acknowledged in its 1999 Carcinogen Risk Assessment Guidelines (RAG), 
that ``in the face of scientific uncertainty, common sense and 
reasonable application of assumptions and policies are essential to 
avoid unrealistic estimates of risk'' (3). CDC believes that a 30-year, 
or even a 10-year, exposure assumption significantly overestimates 
potential exposures by one or more orders of magnitude. For example, 
members of the general public are highly unlikely to be continually 
exposed to sulfur mustard, night and day, for 10 or 30 years. 
Similarly, atmospheric stability, wind speed, and direction are not 
fixed for years on end. No agent reduction is assumed for environmental 
degradations or rainfall that would reduce concentrations. No agent 
reduction is assumed for low temperature environmental conditions where 
mustard agent would not significantly volatilize. No agent reduction is 
assumed for agent dilution beyond the perimeter of a facility. At agent 
storage sites, GPL readings are taken daily at the facility perimeter. 
Levels of agent approaching GPL should be detected within days, not 
years, of occurrence and corrective action would be initiated. 
Historically, agent releases to the environment have been episodic; no 
indication exists that continuous, long term low-level agent releases 
routinely occur.
    CDC's examination of the potential cancer risk associated with 
proposed AELs considered only incremental potential risk. That is, 
historic risk to workers and the public in the vicinity of stockpile 
storage facilities was not examined. This was because each site would 
have to be considered individually regarding amount, nature and age of 
stored mustard items; local spatial, and meteorologic conditions and 
their relation to area demographics; and the nature and capabilities of 
historic storage facility inspection programs. These site-specific 
factors, coupled with a weak quantification of cancer potency (see 
discussion below) of sulfur mustard, suggested limited utility in 
attempting to quantify such potential risk.
    The other major criticism received by CDC regarding carcinogenicity 
analysis pertained to the use of the NAS recommended CPF (2000) based 
upon sulfur mustard relative potency compared with BaP. The NAS 
recommendation was predicated upon oral dosage, not inhalation. CDC 
believed that the other published studies used to support attempts at 
developing numeric estimates of the CPF for sulfur mustard seriously 
lacked merit for this application. Although an averaging estimate 
(i.e., geometric mean) for all the CPFs developed might provide a 
reasonable estimate, CDC believes that a mathematic manipulation of 
questionable numbers in no way ensures that the new number is 
appropriate. Furthermore, CDC believes that without a reasonable basis 
to suggest the estimates used in the averaging method bracket the true 
CPF as applied to humans; CDC should not arbitrarily rely on a number 
developed in this manner.
    CDC agrees with the reviewers that extrapolation between exposure 
routes is undesirable when examining cancer risk. EPA's 1999 Carcinogen 
RAG addresses this issue briefly: ``In the absence of contrary data, 
the qualitative default assumption is that, if the agent is absorbed by 
a route to give an internal dose, it may be carcinogenic by that 
route'' (3). Furthermore, EPA states that, ``For screening or hazard 
ranking, route-to-route extrapolation may be based on assumed 
quantitative comparability as a default, as long as it is reasonable to 
assume absorption by compared routes'' (3). In light of CDC's 
reluctance to use CPF averaged numbers as described above, and in the 
absence of other, better data, CDC recognized that a route-to-route 
extrapolation was needed if the carcinogenicity risk through inhalation 
was to be examined and consequently based its analysis upon the NAS-
recommended potency value.
    CDC believes that the reviewers raise a valid point regarding the 
use of the indexed value as done in the Federal Register proposal. The 
reasonableness of the assumption that both exposure routes result in 
comparable agent absorption is debatable. CDC does not believe strongly 
that such an assumption is valid; consequently, CDC is open to further 
examination of this issue. CDC does not believe that the CPF geometric 
mean offers any demonstrable scientific improvement over the route-to 
route extrapolation originally used in CDC's proposal. The reviewers 
recommend that a range of inhalation cancer slope factors be described 
according to EPA's Carcinogen RAG. CHPPM presented such a range of 
factors in the ``Evaluation of Airborne Exposure Limits for Sulfur 
Mustard: Occupational and General Population Exposure Criteria,'' 
November 2000 and can be referred to by the reader for insight into the 
variability of postulated risk dependent upon a range the exposure 
assumptions and CPFs (1). The CHPPM examination is consistent with 
EPA's guidance. CDC must caution the reader, however, that these 
numeric estimates are tenuous. Oak Ridge National Laboratory's 1993 
discussion of this issue for sulfur mustard carcinogenicity illustrates 
CDC's concerns:
    ``Unfortunately, quantitative human cancer risk estimates are 
impractical because the experimental data from animal studies have 
three large uncertainties:
     Only a few experiments were conducted;
     Many were in a mouse strain that exhibited a 
high genetic susceptibility to spontaneous pulmonary tumors;
     Routes of administration tested and duration of 
follow-up observations are not comparable to the human exposures of 
concern.'' (4)
    In 1991, EPA examined cancer risk estimates that cover the range of 
cancer slope factors presented in the CHPPM document. EPA observed, 
``Depending on the unknown true shape of the dose-response curve at low 
doses, actual risks may be anywhere from this upper bound down to 
zero''(5). Similarly, in the 2003 ATSDR Toxicological Profile for 
Sulfur Mustard, the inhalation cancer effects discussion states, `` * * 
* in no case was the exposure level or duration quantified, and 
therefore, these data are inadequate for deriving dose-response 
    CDC recommends that a better characterization of an appropriate 
cancer slope factor needs to be conducted to set exposure limits. CDC 
is aware of proposed forthcoming animal research by DoD to examine the 
chronic impact of long-term exposure to sulfur mustard. CDC encourages 
this research and the examination of results for possible insights and 
refinement of an estimate of a more accurate CPF.
    5. All four reviewers provided opinions regarding the use of 
uncertainty factors to derive AELs. One reviewer believed that 
rationale was sufficient to reduce the total uncertainty used by the 
National Institute for Occupational Safety and Health (NIOSH) to derive 
the Immediately Dangerous to Life or Health (IDLH) criterion by a 
factor of three, which would result in an increase to a value of 2.0 
mg/m3. Another reviewer wanted to lower the IDLH by a factor 
of two because of limitations of military studies used to derive the 
value. Another reviewer believed strongly that the proposed GPL should 
be reduced by at least an additional factor of 10 to reflect 
uncertainties not adequately represented by either the CHPPM 
examination using the RfC method or the CDC examination using the 
CatReg method. Finally, another reviewer believed that CDC's total 

[[Page 24167]]

factor of 300 used to derive the GPL was appropriate but recommended 
that the uncertainty factor for intrahuman variation be decreased from 
10 to 3 and the data quality factor be increased from 3 to 10. 
Supporting rationale was provided for all these opinions.
    Discussion: Professional judgment is needed in the application of 
uncertainty factors. As discussed in CDC's original support document, 
considerable deliberation is ongoing regarding the use of uncertainty 
factors in risk assessment. No validated or calibrated means exist to 
precisely quantify total uncertainty used in deriving AELs. This was 
why CDC considered not only at the RfC, CatReg, and carcinogenicity 
considerations, but also the risk management aspects of safely managing 
sulfur mustard agent as associated with the demilitarization program.
    The reviewer who recommended the minimal 10-fold decrease in the 
GPL also believed that AELs should be developed independent of risk 
management considerations. CDC agrees that ideally developed AELs 
should be independent of existing risk management conditions. One could 
argue that CDC should ``safe-side'' the AELs by using highest 
uncertainty factors recommended by all reviewers and ignore any 
recommendations for reduction of uncertainty factors. Except for 
compounds exhibiting hormesis, this approach always would be 
theoretically safer than using a number derived using uncertainty 
factors that are not on the most conservative end of the spectrum of 
professional judgment.
    CDC's mission is to enhance public and worker health protection for 
people associated with or living near chemical agent demilitarization 
facilities. CDC believes that real-world risk management must be 
factored into its deliberations. Otherwise, CDC could increase or 
extend actual risk in the real world to minimize theoretical or 
undemonstrated risk. EPA's Carcinogenic RAG noted that, ``While it is 
appropriate to err on the side of protection of health and the 
environment in the face of scientific uncertainty, common sense and 
reasonable application of assumptions and policies are essential to 
avoid unrealistic estimates of risk''(3,6). Furthermore, CDC/NIOSH 
policy for potential occupational carcinogens states that `` * * * 
policy will be the development, whenever possible, of quantitative RELs 
(recommended exposure limits) that are based on human and/or animal 
data, as well as on the consideration of technological feasibility for 
controlling workplace exposures to the REL'' (emphasis added).

Summary and Recommendations

    Although CDC received only 4 sets of comments on the proposed 
mustard AELs, these reviewers clearly tried diligently to represent 
their perspectives and concerns. Three sets of comments focused 
primarily upon the process used to develop the proposed AELs, and the 
fourth focused primarily on the practical implications of the proposed 
values. In addition to the solicited comments described above, CDC had 
the original proposal reviewed by other government and professional 
health risk assessment personnel. With the exception of one reviewer, 
the CDC approach to developing AELs in concert with ongoing risk 
management provisions of the chemical demilitarization program was not 
    The examination of the carcinogenicity issue is problematic in that 
CDC believes that a numeric estimation of a cancer slope factor for 
mustard is not well supported. The CHPPM review of this issue, through 
the evaluation of the range of attempts at quantifying upper bound 
cancer risk from exposure to sulfur mustard, has been referenced herein 
to provide the reader with that perspective; however, CDC cannot say 
with confidence that the numeric range of slope factors is likely to 
provide a reasonable estimate of the true carcinogenic potency of this 
    Because of the uncertainties discussed above, especially the 
characterization of cancer potency of sulfur mustard, CDC has decided 
to issue its recommended AELs as interim values pending better 
understanding of the CPF for this agent. CDC believes that for 
noncancer effects, the recommended AELs protect worker and public 
    Regarding the implied carcinogenicity risk, CDC believes that the 
strong risk management provisions, such as engineering and 
administrative controls within demilitarization facilities, extensive 
low-level air monitoring, and the previously discussed mitigating 
factors, minimize cancer risk at the interim AELs.
    In summary, CDC recommends the following:
     Defer recommending a cancer potency factor until 
better data are available.
     Redesignate the ceiling-5M value as a 15-minute 
STEL, limited to one occurrence per day; CDC encourages shortest 
practicable analytic cycle times.
     Apply the U.S. Army CHPPM-derived 8-hour WPL for 
workplace; retain GPL as proposed for use in protecting the general 
     Implement the recommended AELs as interim 
values, to go into effect on July 1, 2005; values to remain interim 
until better cancer potency characterization is available or research 
data indicate the need for revision.
     Continue to recommend rigorous risk management 
analysis and practice as has been associated with the chemical agent 
demilitarization program practice.
     Given the uncertainty in the risk assessment 
regarding cancer potency, reduced exposures to sulfur mustard to the 
lowest practicable level.
    Table 1 below contains the numeric values for the interim 
recommended AELs.

    Dated: April 27, 2004.
James D. Seligman,
Associate Director for Program Services, Centers for Disease Control 
and Prevention.

                           Table 1.--CDC Recommended Interim Airborne Exposure Limits*
  [All values expressed as mg/m\3\ in air with concentration x time [Ct -- mg-min/m\3\] values in parentheses]
     Sulfur mustard (H, HD           General       Worker population   exposure limit   Immediately dangerous to
      [dagger]) criteria         population limit        limit            [Dagger]          life or health 
Exposure Level................  0.00002 (0.01)...  0.0004 (0.19)....  0.003 (<=0.04)..  0.7 (<=21).
Averaging Time................  12 hours.........  8 hours..........  <=15 minutes....  <=30 minutes.
Recommended Monitoring Method.  Historic   ....  Historic    or   Near-real-time..  Near-real-time.
* Although CDC does not specifically recommend additional reduction factors for statistical assurance of action
  at the exposure limit, exposures to sulfur mustard should be minimized given the uncertainties in risk
  assessment, particularly as related to characterizing carcinogenic potency.

[[Page 24168]]

[dagger] The toxicity data for agent T is inadequate for setting exposure limits. The very low vapor pressure
  for agent T precludes it as a vapor hazard under normal ambient conditions. For sulfur mustard and T mixtures,
  air monitoring for sulfur mustard alone should be sufficient under most circumstances to prevent exposure to
[Dagger] To be evaluated with near-real-time instrument using shortest practicable analytic cycle time. No more
  than one exposure per work-shift.
  The 30-minute period is not meant to imply that workers should stay in the work environment any longer than
  necessary; in fact, they should make every effort to exit immediately. IDLH conditions require highly reliable
  dermal and respiratory protection.
   Historic monitoring typically is used for time-weighted average (TWA) monitoring where the sample analyzed
  represents an extended time period, e.g., 8 or 12 hours. Results are not known until laboratory analysis is
  completed after the sampling event. AELs using historic monitoring are set at levels at which health effects
  are not expected to occur for most workers. Exposures above the WPL-8, but below the STEL, likewise are not
  expected to result in significant health effects unless such exposures occur continuously for long periods.


1. U.S Army. Evaluation of Airborne Exposure Limits for Sulfur 
Mustard: Occupational and General Population Exposure Criteria--47-
EM-3767-00. Aberdeen Proving Ground, MD: U.S. Army Center for Health 
Promotion and Preventive Medicine, November 2000.
2. U.S. Department of Health and Human Services: Toxicological 
Profile for Sulfur Mustard (Mustard Gas). Atlanta, GA: Agency for 
Toxic Substances and Disease Registry, September 2003.
3. U.S. Environmental Protection Agency: Draft Guidelines for 
Carcinogen Risk Assessment. Washington, DC: U.S. Environmental 
Protection Agency, July 1999; Publication No. NCEA-F-0644.
4. Nicholson W, Watson A. Risk assessment considerations for sulfur 
mustard. In: Pechura CM, Rall DP (eds.) Veterans at Risk: The Health 
Effects of Mustard Gas and Lewisite, Washington, DC: National 
Academy Press, 1993: 390-8.
5. U.S. Environmental Protection Agency: Upper-Bound Quantitative 
Cancer Estimate for Populations Adjacent to Sulfur Mustard 
Incineration Facilities. Washington, DC: U.S. Environmental 
Protection Agency, Office of Research and Development, July 1991, 
Publication No. EPA/600/8-91/053.
6. U.S. Environmental Protection Agency: Policy for Risk 
Characterization. Memorandum of Carol M. Browner, Washington, DC: 
U.S. Environmental Protection Agency, March 21, 1995.

[FR Doc. 04-9946 Filed 4-30-04; 8:45 am]