ECDIC-1998-48
FEDERAL INSECTICIDE, FUNGICIDE, AND RODENTICIDE ACT (FIFRA)
'
GOOD-.LABORATORY PRACTICE STANDARDS (GLPS)
: QUESTIONS AND ANSWERS
Prepared by .the
Pesticides Enforcement Policy Branch
Policy and Grants Division
Office of Compliance Monitoring
Office of Prevention, Pesticides, and Toxic Substances
U.S. Environmental Protection Agency
May 12, 1992
-------
FIFRA GLP Q's & A's
May 12. 1992
Page 1 of 14
INTRODUCTION
On August 17, 1989, EPA published in the Federal Register revisions to the
Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) Good Laboratory Practice
standards (GLPS) (54 FR 34052). This revision included changes that the Food and
Dnig Administration made to its GLPS (September 4, 1987; 52 FR 33768) and expanded
the scope of the regulations to include data submissions which had previously not been
under GLPS. The expansion-of GLPS to include field studies has brought many facilities
under GLPS for the first time while also making the standards applicable to entirely
different..types of testing, environments than had previously been the case.
Since the publication of the revised rule in- 1989, EPA has received many
' .questions from persons who wish clarification regarding the applicability of the rule to
their activities. These questions have ranged from simply asking whether the work they
are doing-.is .required to comply, to technical questions regarding how the standards
should "be applied In. the context of field as opposed to laboratory studies. Many written
replies have been.made to persons who have submitted specific questions in writing to
EPA, Copies of specific correspondence have been provided upon request.
Notwithstanding, the-correspondence file is of limited usefulness to other persons
since the issues addressed are often specific to a particular situation. There have been .
requests for a general guidance document regarding EPA's FIFRA GLP policy. The
following questions and answers have been prepared by the Policy and. Grants Division ot
the Office of Compliance Monitoring to serve as official written policy for -the regulated
community. . . .
-------
FIFRA GLP Q's &. A's
May 12, 199:
Page 2 of 14
QUESTIONS AND ANSWERS
APPLICABILITY
1. What is the applicability of GLPS to work in progress at the time that the rule
became effective?
The GLPS apply to all study-related work which is performed on or after the .
effective date of the rule. Studies in progress must be in compliance with GLPS
from the effective date onward. A statement of compliance or non-compliance
must accompany .the final study report for such a study. This statement must
either- (1) state that the study was in compliance with GLPS, (2) describe in detail
how it did not comply with GLPS, or (3) state that the submitter did not sponsor
or conduct the study and does not know its compliance status. The statement .
must account for compliance or deviations with both the previous GLP. rule
(effective 1984), and the current rule (effective 1989),.as applicable! .
2. If a study was in progress on October 16, 1989, must it have a protocol? What
parts of the study would the protocol address?
All portions of the study performed on or after the effective date must be
performed according to a written protocol as provided at 40 CFR 160.120. That.
protocol need only address those parts of the study performed on or after the
effective date. Please note that if a study was subject to the 1984- GLPS, a
..protocol was required for all parts of the study conducted after the effective date
. of that rule. The compliance statement submitted with that study's report must
specify in detail those study activities which were not performed in accordance
with GLPS.
3. Current reregistration procedures involve submission of data that resulted from
studies, performed prior to the effective date of GLPS. Do GLPS apply to such
data, and if so, how?
Any data presently submitted in support of a pesticide research or marketing
permit must be accompanied by a true and correct compliance statement as
described at 40 CFR 160.12 regardless of when the study was performed.
Therefore, .data submitted to meet reregistration requirements are required to be
accompanied with a true and correct compliance statement informing EPA in
detail of all differences between the practices used in the study and those required
by GLPS. It is.not unlawful to.truthfully admit that studies supporting such-
-------
FIFRA GLP Q's & A's
May 12, 1992
Page 3 of 14
submissions did not comply with GLPS, nor-would such an admission necessarily
lead to rejection of the data. The compliance statement'will help the Agency to
determine the reliability of the data based on current data requirements. .Note
that such an admission may nevertheless result in an enforcement action if they
indicate that an unlawful act has occurred. For example, other regulations, i.e.,
books and records 'as stated at 40 CFR 169.2(k), require retention of raw data
generated in support of registered pesticides prior to the effective date of GLPS.
. Admitting to destruction of records would not exclude the Agency from taking
enforcement actions for. the books and records violation.
4. Do GLPS apply to data used to support tolerance petitions?
Yes. The scope of the regulations as stated at 40 CFR 160.1 require that studies
conducted to develop data, pursuant, to sections 408 and 409 of the Federal Food,
. Drug, and-Cosmetic Act be performed in accordance with GLPS.
5. Are studies conducted under the Interregional Research Project Number 4 (IR-4)
program to support the registration of minor uses subject to .the GLPS?
-. Yes. .''''' '
6. Do GLPS apply to weather data and soil analysis data?
Any data which are collected as pan of a study listed in 40 CFR 160.1 must be
collected according to GLPS. This includes weather data and soil analyses which
are collected as part of a larger study which must comply.with GLPS. If non-study
data such as local weather data are cited in a study report, and the study report
clearly indicates that such data were not gathered as part of the study, GLPS
would not apply to such data.
7. What applicability do GLPS have when State, Federal, or independent
laboratories are used to provide soil or weather data for GLP studies?
GLPS are .applicable in.such circumstances if such data are gathered as part of a
FIFRA study. Only where such data are gathered independently of the study, and.
the study report clearly indicates that such data were not gathered as part of the
study, would GLPS not apply.
-------
FIFRA GLP Q's. & A's
May 12, 1992
Page 4 of 14
DEFINITIONS -
8. Wfll EPA issue separate GLP standards for field testing as opposed to laboratory
testing?
. The expansion of GLPS to cover field studies was based on the need to assure
identical standards for all data submitted to EPA under FIFRA, and on the
determination that the GLPS are technically general enough to cover virtually any
type of research environment. EPA does not intend to issue separate standards.
9. Can an experiment be divided into more than one study, based on where- or when
the work is performed, or the phase of the experimental work?
Under GLPS, the .term "study" refers to an experiment to determine or predict the
".effects or characteristics of a. test substance. EPA considers a study to be
composed of all of the necessary elements of research which are performed in
order to obtain the reported results. If the elements of .research consist of several
.phases of work which must be taken in the context of each other to get
meaningful results, they are all considered to be .elements of the same study. An
example of this would be where one laboratory treats a test system with a test
substance and sends the treated test system to another laboratory for analysis.
If the experiment involves treatment of test systems in several different locations,
the experiment may be considered to be composed of either one study
encompassing all locations or several studies each involving one or more locations.
In the latter case, however, it would be necessary, that each separate study stand
entirely by itself, i.e., meet all of the criteria of a study. There would have to be
separate compliance statements for each, separate tracking on master schedules,
separate quality assurance inspections, etc. Each study would have to have a
study director (and only one study director), although it may be possible for the
same study director to oversee several of such studies at the same time. Finally,
where several studies are compiled for submission, the submission must include
true and correct compliance statements for each study involved in the submission.
10. What is EPA's formal policy on certifying copies of raw data? Must each page be
signed and dated?
. EPA stated in the preamble to the August 17, 1989 rule (54 FR 34066) that
acceptable alternatives to signing and dating each page may be devised ancj
incorporated into standard operating procedures. EPA did not further elaborate
in.order to allow, each testing facility flexibility in implementing SOPs that would-
-------
FIFRA GLP Q's & A's
May 12, 1992
Page 5 of 14
provide adequate assurances within its facilities. Note that EPA may inspect the
original records, which must be maintained by the registrant as provided at 40
CFR I69.2(k), to assure that they have been kept and that the copies are correct.
11. Is it permissible to discard original raw data worksheets after exact copies have
been made? .
Destruction of original raw data is prohibited. The registrant is responsible, for
maintaining all original raw data as specified at 40 .CFR 169.2(k). Copies of data
may be used to assure compliance with GLPS at the level of the testing facility,
but EPA requires that the registrant maintain all original data that support a
study. ' . :
12. What type of sponsor-testing facility communication is considered to be raw data
which must be archived at the end of the study?
All records of sponsor-testing facility communication which occur as part of the
activities of a study are."considered to be raw data, as defined at 40 CFR 160.3.
This includes memoranda, letters, and records of telephone conversations which
occur during the course of the study. Communication conducted prior to the
study (i.e., before the protocol is signed) or following the completion of the study
(i.e., after the report is signed) would not normally be considered to be raw data.
Note that certain records not specific to a particular study which are generated
when the study is not in progress still need to be retained to prove that study's.
compliance with GLPS. Examples include records of a sponsor's notifying a
facility of the need to comply with GLPS as required at 40 CFR 160.10, and
records of facility documents such as. standard operating procedures.
STUDY DIRECTOR
13. Many field studies involve more than one technical phase, each involving different
personnel and different methodologies, often by different contractors. Concern
has. been raised over the difficulty for a single individual to physically oversee all
phases and to be expert in all techniques involved. Within the same study, is it
acceptable to assign a different study director to different phases?
No. Each study must have a single study director who represents the single source
of study control. This is explicitly stated in the GLPS at 40 CFR 160.33. A single
point of control is necessary to the integrity of the study and to avoid the potemi;ii
for conflicting instructions and confusion in study implementation.
-------
FIFRA GLP Q's & A's
May 12, 1992
Page 6 of 14
14. If there can only be one study director assigned to a study, is it acceptable to
assign "field directors" and "analytical directors" to manage the work which
: involves different phases and/or locations?
The assignment of responsibility for the study to the study director need not
interfere with ordinary delegation of authority necessary for the performance of
,... study duties. Any authority accepted by persons other than the study director
does not reduce the study director's overall responsibility for the study.
QUALITY ASSURANCE UNITS
15. Is it acceptable to inspect study-related procedures at a time other than when the
study is ongoing? .
The GLPS state at 40 CFR 16Q.35(a) that a testing facility shall have a Quality
^Assurance Unit (QAU) that shall monitor each study to assure management that
the facilities, equipment, personnel, methods, practices, records, and controls are
in conformance with the GLPS; The GLPS further, state at 40 CFR 160.35(b)(3)
that the QAU shall inspect each study at intervals adequate to ensure the integrity
. of the study. '
Clearly, the QAU must conduct inspections adequate to provide the assurances
required at 40 CFR 160.35(a) and, in the course of so doing, must inspect each
study at least once. All parameters must be verified adequate for each site, but it
is acceptable to use inspections conducted during other studies to provide
necessary assurances. It is also acceptable to use inspections conducted when no
study is in progress to assure that methods, personnel, etc. at a particular site are
in conformance with GLPS. However, acceptability of such inspections is
contingent on assuring that the facilities, personnel, methods, etc., which are .
inspected are representative of those used in the study. Note that it is necessary
to reinspect facilities periodically to account for changes in personnel, equipment,
etc. Finally, no matter how complete QAU inspectional coverage is regarding the
sites involved in a study, it is still necessary to conduct at least one inspection of
study activities while the study is in progress.
-------
FIFRA GLP Q's & A's
May 12, 1992
Page 7 of 14-
16. What would constitute adequate inspection of the ongoing study? Would an audit
of the protocol or of data records be adequate?
At least one inspection must be conducted while the study is in progress. Under
GLPS, the QAU monitoring of protocols, data records, or other documentation
phases of a study are important just as is directly observing the experimental
phase'of the study. However, the GLPS state at 40 CFR I60.35(b)(3) that
inspections must be done at intervals adequate to ensure.the integrity of the study,
and further, at 40 CFR 160.35(b)(4), that periodic status reports noting problems
and corrective actions be submitted to management.
An. audit of a study protocol would be of very limited utility since the subsequent
reporting would be to management which, in all likelihood, has already reviewed
. < the protocol. Data record audits would also be of very limited utility since they
may occur after all experimental work, is completedin short, too late for any
corrective actions to be taken. This problem also applies to protocol audits
conducted after the experimental phase is completed. Thus, reliance solely on
such types of audits wouid not meet the GLP requirements as stated at 40 CFR
160.35.-, - ----- ' . -
FACILITIES' , -
17. Is it permissible to store mixed feeds containing the test substance in the same
room with the test system during feeding studies?
As discussed at 40 CFR 160.47(b) test substance mixture storage areas must be
stored in separate areas from the areas where test systems are kept However,
working quantities of test substance mixtures need not be stored in separate
rooms from test systems. Separate areas within the same room may be designated
for test substance mixture, storage and test systems as long as the separation is
adequate to preserve the integrity of the study and the identity, strength, purity
and stability of the mixture.
-------
FIFRAGLP.Q's& A's
May 12, 1992
Page 8 of 14
TEST. CONTROL AND REFERENCE SUBSTANCE CHARACTERIZATION
18. Do characterization requirements at 40 CFR 160.105 apply to analytical
standards?
Analytical standards are considered to be reference substances and are subject to
all GLP standards that apply to reference substances, including characterization.
-^
19. Can data developed by the supplier of the standard be accepted? If not, can it be
used on an "interim" basis untfl the standard is adequately characterized?
Information developed by a supplier can be used to support characterization .
requirements, but the compliance statement for the overall' study must state
whether such data were developed under GLPS. Any data not developed under
GLPS may be rejected by the Agency. Analyses must be performed to
characterize the reference .substance before.it is used. In the rase that a standard .
is used before it is analyzed, this is a"violation of 40 CFR 160.105(a), which
requires such determinations to be made before the standard is used in the study.
"'*'. *
20. What documentation would apply to standards?
Full characterization information as stated at 40 CFR 160.105 is required of
standards. This section requires that any information that is appropriate for.
defining the standard, including identity, strength, purity, or composition, shall be
determined for each batch before it is used. In the case of an analytical standard,
for example, it is necessary to obtain analysis data documenting the identity,
strength, and purity, for each batch. A labeled assay value, in and of itself, is
:. insufficient
FEST SUBSTANCE STORAGE CONTAINERS - '
11. Is it necessary to retain test substance storage containers for the duration of a
field study?
Yes, as provided at 40 CFR 160.105(c), storage containers for test substances shall
be assigned for the duration of a study. This requirement is necessary to assure
that test substances are stored in proper containers, and that the containers that
are used can be accounted for during the study, At any time during .the study, it
must be possible to examine the containers to assure that this standard is met.
However, requests for waivers involving large numbers of containers or safety
-------
FIFRA GLP Q's & A's
May 12, 1992
Page 9 of 14
concerns may be made to the Director, Policy and Grants Division (see question
#23). " -
22. If a large number of containers are involved in a study and/or unusual safety
problems are caused by the storage of such containers, is there any alternative to
storage?
Yes, but only if written permission is obtained, from the Director, Policy and
Grants Division (see question # 23). The written letter authorizing disposal of
containers will impose certain requirements that will ensure that the iritent of the
GLP standards are met. '
23. How does one obtain such permission?
A request for permission must be submitted in writing to the Director, Policy and
Grants Division, Office of Compliance Monitoring (EN-342), U.S. Environmental
Protection Agency, 401 M Street, SW, Washington, DC 20460. The request must
identify the study for which permission is requested, the testing facility, the nature
and quantity of containers involved, and the time and locations) of the study.
The request should also identify any special storage burdens orsafety hazards
which retention of the containers may pose. ,
24. What types of conditions would be imposed by EPA in granting such permission?
EPA will request that sufficient'documentation be available to assure that any
containers-which have been used for test substance storage during the course of a
study are thoroughly accounted for from the time of receipt to disposal This
documentation would generally include such items as bills of lading, inventory
records, receipts, use logs, and any other supportive records. In addition, the
letter will stipulate that the Director of the Laboratory Data Integrity Assurance
Division of OCM be notified of the location of such records in order that they be
available for inspection.
25. Can "generic" permission be obtained to cover multiple studies and/or test
substances?
No. Each case will be evaluated individually. However, more than one study
and/or test substance may be included in given request, as long as each study and
test substance is specifically identified. v
-------
FIFRA GLP Q's &. As
May 12, 199"
Page 10 of l-
PROTOCOLS
26. Can a "generic protocol" be used for obtaining sponsor approval?
The GLPS require that the protocol be approved by the .sponsor, and the date of
approval must be included with the protocol; however the GLPS also provide
flexibility in how this approval is obtained. A "generic protocol" approach may be
acceptable for obtaining sponsor approval of certain protocol elements. .In such a
case, the testing facility which is drafting the protocol for a study would, only need
to obtain approval of those elements which were not included in the generic
protocol. Please note that since the GLPS require protocols to include certain
information that would not be included in a generic protocol, such as the test
substance or the proposed start and termination dates, it would still be necessary
to obtain sponsor approval .for such information in.addition to the approval of the
generic protocol. . : .
27. What records of seeds or transplants of crops or plants used in 6eld studies must
be maintained? . " ___'"".'..._---
Where crops or plants are-the test system or a component.of the test system, all
GLP standards relating to test system records are applicable. These include
protocol provisions given at 40 CFR 160.120(a)(6) and (7), as applicable.
Included, for example, would be the source of the. test system supply, species,
method of identification, etc. Lot numbers of .seeds, brand names, and other '
information uniquely identifying the test system would be relevant.
REPORTING * .
28. The GLPS at 40 CFR 160.185(a)(12) require that signed and dated reports of
each scientist or other professional in the study be included in the final report.
Can these reports be combined into one report, with all of the scientists and
professionals dating and signing that report?
This requirement is intended to ensure that all information related to the study is
included in the final report. Specifically, when individual scientists findings are
.part of the study effort, they are required to be included separately. Combined
reports may in effect be consensus documents, and that would defeat the purpose
of this requirement. Note that this requirement is not intended to require,
separate reports of all scientists participating in a study if such scientists are not.
-------
FIFRA GLP Q's & A's
May 12, 1992
Page 11 of. 14
in fact, providing individual findings or opinions. For examplerpathologisrs
reports are considered to be separate findings which must be reported separately.
ARCHIVES .'
29. The GLPS state that the study director must assure that raw data are transferred
to archives during or at the close of the study. Is there a "grace" period allowed
after the end of the study to allow this to be done?
Under GLPS, the study director is required to assure that all raw data,
documentation, the protocol, specimens, and final reports are transferred to the
archives during or at the close of the study (40 CFR 160.33(f)). Thus, there is no
grace period. The study director must comply with this requirement prior to
signing the compliance statement. This ensures that data are fully accounted for
at the completion of the study. . .
30. How does EPA define "close of study" in regard to archiving?--
The term "at the close of the study" is strictly interpreted to mean that point of
time at which the study director signs the final study report. The act of signing
. the final report is one of assurance by the study director that the report is a true
representation of the data that support the report... At or prior to the time that
the study report is signed, the study director must pass control of the raw data to
the archives where their integrity will be maintained. Any delay in the transfer of
data beyond the close of the study creates a lapse between the time that the study
director assures that the raw data support the study report and the time that the
data are secured from damage, misuse, or loss. :
31. Given that data must be transferred to archives at the close of the study, is it
possible to use temporary archives prior to transfer to a central archive?
There is flexibility in the location of the archives of raw data and specimens. At
40 CFR 160.190(b), the GLPS state that retention of records at alternate locations
is acceptable, provided that there is specific reference to those locations in the
archives. Such off-location archives must still meet the full requirements of 40
CFR 160.190. Whether records are archived at the registrant's facility, at a
contractor's central location, or at separate contractors' locations, the study
director must assure that all raw data and specimens have been archived before
the study report is signed. If the study director cannot assure that .records at a
particular location are archived correctly, he should not sign a compliance
statement that indicates that this standard has been met. Note that, for the
-------
FIFRA GLP Q's & A's
May 12, 1992
Page 12 of 14
purpose of complying with GLPS, true copies may be archived" at the ciose of the
study. The_ original records will have to be maintained as well but need not be
archived at the end of the study if this is impractical, for example where the
original data constitutes a facility record shared by other studies still in progress at
the close of the study.
32. Is it necessary to retain frozen tissue samples in archives, or may these be
discarded after quality assurance verification?
Under FIFRA GLPS, 40 CFR 160.195, frozen tissue samples are required to be
retained in archives, and there are no specific allowances for their being discarded
as there are for "specimens obtained from mutagenicity tests, specimens of soil,
water, and plants, and wet specimens of blood, urine,, feces, and-biological Quids."
The GLPS do not require specially prepared material to be retained beyond the
period that it affords evaluation if such, material is relatively fragile and differs
markedly in stability or quality during' storage. EPA does noj^ believe, that this is
the case for many types of frozen tissues. The reason that tissues are frozen is to
retain their utility for evaluation. Please note that, as provided at 40 CFR
160.195(h), non-documentary material such as samples and specimens may be
discarded after EPA has notified the sponsor or testing facility in writing that
retention is no longer required. .
33. Must field notebooks be archived during or at the close of a study?
If a notebook contains raw data, the notebook or the raw data must be archived
,. at the close of the study. Note that the registrant is responsible for the original
records under 40 CFR 169.2(k) and section 8 of FIFRA, so it is inadvisable to
enter raw data for studies related to different registrations in the same bound
notebook.
34. Must analytical preparations (e.g., scintillation vials and solutions) be archived?
Such preparations need not be retained beyond the period that they afford
evaluation, as stated at 40 CFR 160.195(c). Generally, samples prepared for
analysis have limited utility beyond the time of analysis and can be discarded.
35. How long must sou", water and plant specimens be retained?
These need only be retained until the QAU has verified that their disposal( will nut
jeopardize the integrity of the study, as provided at 40 CFR 160.190(a) and
160.195(c). Please note that there may be study-specific sample retention
-------
FIFRA GLP Q's & A's
May 12. 1992
Page 13 of 14
requirements in addition to and independent of GLP requirements. Failure to
retain such -samples may result in rejection of data by EPA or enforcement actions
independently of whether a GLP violation has occurred.
GLP VIOLATIONS
36. Can EPA assess penalties for GLP violations?
Yes. FIFRA section 14 states the EPA's authority to assess penalties for
violations of the Act.
37. What are the possible violations under the statute?
. Violations of GLPS may constitute unlawful acts, under FIFRA. Under section
12(a)(2)(M) it is unlawful to knowingly falsify all or pan of any application for
registration, application for experimental use permit, any information submitted to
the Administrator pursuant to section 7, any records required to be maintained
pursuant to this Act, any report filed under this Act, of any information marked as
confidential and submitted to the Administrator under any provision of this Act to
be submitted to EPA or of records required to be maintained. Under section
12(a)(2)(Q) of FIFRA it is unlawful to falsify all or pan of any information
relating to the testing of any pesticide (or any ingredient, metabolite, or
degradation product thereof), including the nature of any protocol, procedure,
substance, organism, or equipment used, observation made, or conclusion or
opinion formed, submitted to the Administrator, or that the person knows will be
furnished to the Administrator, or will become a part of any records required to
be maintained by this Act. Under section 12(a)(2)(R) of FIFRA it is unlawful to
submit to the Administrator data known to be false in support of a registration.
Finally, it is unlawful under FIFRA section 12(a)(2)(B)(i) of FEFRA to refuse to
prepare, maintain or submit any records required by or under sections 5, 7, 8, 11,.
or 19. -
38. What are the maTimuTn penalties that can be imposed?
Section 14(a) of FIFRA provides for maximum civil penalties of not more than
$5000 per offense for violations of the Act by registrants, commercial applicators,
wholesalers, dealers, retailers, or other distributors, and of not more than $1000
per offense for other persons. For knowing violations of the Act, FIFRA section
14(b) provides for maximum criminal penalties of not more than $50,000 and/or 1
-------
FIFRA GLP Q's & A's
May 12, 1992
Page 14 of 14
year imprisonment for producers, registrants, or applicants for registration and of
not more than S25,000 and/or 1 year imprisonment for other knowing violators.
39. Wfll civfl or criminal penalties be imposed for all GLP violations?
No. Section 9(c)(3) of FIFRA allows a written notice of warning to be issued for
a minor violation, if such warning is determined to be adequate to serve the public
interest. Section 14(a)(4) of the Act further provides that in determining the size
of a penalty EPA may issue a warning in the case that a violation occurred despite
exercise of due caution or did not cause significant harm to health or the
environment. Finally, section 14(a)(2) of FIFRA provides that persons other than
registrants, commercial applicators, wholesalers, dealers, retailers or other
distributors, who violate any provision of the Act may be assessed a civil penalty
only subsequent to receiving a written warning for a prior violation. Thus, persons
who only perform testing and are not engaged in the distribution and sale of
pesticides will not be assessed civil penalties for. their first offense. This does nor
extend to criminal penalties as described, at section 14(b)(2) of FIFRA.
40. Can EPA reject studies not conducted in accordance with GLPS?
Yes. The regulations specifically provide for this at 40 CFR 160.17(a), which
states that "EPA may refuse to consider reliable ... any data from a study which
[is] not conducted in accordance with [GLPS]." GLP violations associated with a
study submitted to EPA may also result in enforcement actions whether or not a
study is rejected.
------- |