[Federal Register Volume 68, Number 121 (Tuesday, June 24, 2003)]
[Rules and Regulations]
[Pages 37620-37658]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 03-14425]



[[Page 37619]]

-----------------------------------------------------------------------

Part II





Department of Transportation





-----------------------------------------------------------------------



National Highway Traffic Safety Administration



-----------------------------------------------------------------------



49 CFR Part 571



Federal Motor Vehicle Safety Standards; Child Restraint Systems; Final 
Rule

  Federal Register / Vol. 68, No. 121 / Tuesday, June 24, 2003 / Rules 
and Regulations  

[[Page 37620]]


-----------------------------------------------------------------------

DEPARTMENT OF TRANSPORTATION

National Highway Traffic Safety Administration

49 CFR Part 571

[Docket No. NHTSA-03-15351]
RIN 2127-AI34


Federal Motor Vehicle Safety Standards; Child Restraint Systems

AGENCY: National Highway Traffic Safety Administration (NHTSA), 
Department of Transportation.

ACTION: Final rule.

-----------------------------------------------------------------------

SUMMARY: This document makes a number of revisions to the Federal 
safety standard for child restraint systems, including amendments for 
incorporating improved test dummies and updated procedures used to test 
child restraints and extension of the standard to apply it to child 
restraints recommended for use by children up to 65 pounds (30 
kilograms). This action strengthens the technical underpinnings of the 
standard and ensures that a firmer foundation is laid for possible 
technical improvements in the future. Child restraints will be tested 
using the most advanced test dummies available today and tested to 
conditions representing current model vehicles. This final rule 
fulfills the mandate of the Transportation Recall Enhancement, 
Accountability and Documentation Act of 2000 that the agency undertake 
rulemaking on the safety of child restraints.

DATES: The amendments made in this rule are effective December 22, 
2003. The incorporation by reference of certain publications listed in 
the rule is approved by the Director of the Federal Register as of 
December 22, 2003. If you wish to petition for reconsideration of this 
rule, your petition must be received by August 8, 2003.

ADDRESSES: If you wish to petition for reconsideration of this rule, 
you should refer in your petition to the docket number of this document 
and submit your petition to: Administrator, Room 5220, National Highway 
Traffic Safety Administration, 400 Seventh Street SW., Washington, DC 
20590.

FOR FURTHER INFORMATION CONTACT: For non-legal issues, you may call 
Mike Huntley of the NHTSA Office of Crashworthiness Standards, at 202-
366-0029.
    For legal issues, you may call Deirdre Fujita of the NHTSA Office 
of Chief Counsel, at 202-366-2992.
    You may send mail to both of these officials at the National 
Highway Traffic Safety Administration, 400 Seventh St., SW., 
Washington, DC 20590.

SUPPLEMENTARY INFORMATION:

Table of Contents

I. Executive Summary
II. Background
III. The TREAD Act
IV. Responsible Regulation
V. Response to the TREAD Act
    a. NPRM For This Final Rule
    b. ANPRM On Side Impact Protection
    c. TREAD Programs on Labeling and Consumer Information
VI. Post-NPRM Testing
VII. Summary of Comments on the NPRM
    a. General Comments on the Proposals
    b. Updated Bench Seat
    c. Crash Pulse
    d. New Dummies
    e. Application of the Standard
    f. Injury Criteria
VIII. Amendments
    a. Updated Bench Seat
    1. Post-NPRM Test Program
    i. Seat Back Rotation
    ii. HIC Measurements
    iii. Chest Acceleration
    iv. Head Excursion
    v. Knee Excursion
    vi. Summary of PAX Testing
    2. Response to Comments
    i. Seat Back and Cushion Angles
    ii. Belt Systems On The Standard Seat Assembly
    iii. Fixed Seat Back
    iv. Future Work
    b. Crash Pulse
    1. On Widening The Corridor
    2. Increase Pulse Severity
    3. Decrease Pulse Severity
    c. New Dummies
    1. Post-NPRM Test Program
    i. Tests With The 3-Year-Old Dummies
    ii. Tests With The 6-Year-Old Dummies
    2. Commenters Generally Supportive
    i. Hybrid III 6-Year-Old Dummy
    ii. Weighted 6-Year-Old Dummy
    3. Specific Issues Relating To The Use Of The New Dummies In 
Standard No. 213
    i. Seat Back Height Requirement
    ii. Padding Requirement
    d. Application of the Standard
    e. Injury Criteria
    1. Post-NPRM Testing
    i. JPMA
    ii. NHTSA Series I and II
    iii. Results of JPMA and NHTSA Series I and II
    iv. NHTSA Series III
    2. Comments and Conclusions
    i. Head Injury Criterion (HIC)
    ii. Chest Injury Criteria
    iii. Neck
    f. Leadtime
IX. Regulatory Notices and Analyses

I. Executive Summary

    This document makes a number of revisions to Federal Motor Vehicle 
Safety Standard (FMVSS) No. 213, ``Child Restraint Systems'' (49 CFR 
571.213). The revisions incorporate four elements into the standard: 
(a) An updated bench seat used to dynamically test add-on child 
restraint systems; (b) a sled pulse that provides a wider test 
corridor; (c) improved child test dummies; and (d) expanded 
applicability to child restraint systems recommended for use by 
children weighing up to 65 pounds. This action strengthens the 
technical underpinnings of the standard and ensures that a firmer 
foundation is laid for possible technical improvements in the future. 
Child restraints will be tested using the most advanced test dummies 
available today and tested to conditions representing current model 
vehicles. This final rule does not adopt the scaled injury criteria 
developed for the occupant protection standard (FMVSS No. 208), except 
that the time interval used to calculate the head injury criterion is 
amended from an unlimited time interval to 36 milliseconds.
    This final rule fulfills the mandate in the Transportation Recall 
Enhancement, Accountability and Documentation Act (the TREAD Act) 
(November 1, 2000, Pub. L. 106-414, 114 Stat. 1800) to initiate a 
rulemaking for the purpose of improving the safety of child 
restraints.\1\
---------------------------------------------------------------------------

    \1\ It also follows up on the agency's announcement in its 
November 2000 Draft Child Restraint Systems Safety Plan (Docket 
NHTSA-7938) that the agency will be undertaking rulemaking on these 
and other elements of Standard No. 213 (65 FR 70687; November 27, 
2000).
---------------------------------------------------------------------------

    Section 14(a) of the TREAD Act mandated that the agency ``initiate 
a rulemaking for the purpose of improving the safety of child 
restraints, including minimizing head injuries from side impact 
collisions.'' Section 14(b) identified specific elements that the 
agency must consider in its rulemaking. The Act gave the agency 
substantial discretion over the decision whether to issue a final rule 
on the specific elements. Section 14(c) specified that if the agency 
does not incorporate any element described in Sec.  14(b) in a final 
rule, the agency shall explain in a report to Congress the reasons for 
not incorporating the element in a final rule.
    In response to Section 14, the agency examined possible ways of 
revising and updating its child restraint standard. Today's final rule 
is substantially based on a combination of pre- and post-TREAD Act 
agency activities, including research studies of child restraints and 
dummies by NHTSA following issuance of the NPRM. This final rule was 
also developed based on extensive information provided by comments to 
the NPRM. Several factors relating to child restraint performance and 
use in this country guided the agency in its decision-making on this 
rulemaking, in addition to the statutory mandates

[[Page 37621]]

governing the agency's rulemaking activities. These factors are 
outlined in Section IV of this preamble.
    The agency also issued an advance notice of proposed rulemaking 
(ANPRM) published concurrently with the NPRM, in which comments were 
sought on the agency's work on developing a possible side impact 
protection standard for child restraint systems. This advanced notice 
is discussed in Section V of today's preamble. The ANPRM announced that 
the agency had conducted extensive testing and analysis over the year 
proceeding the ANPRM to develop a possible side impact protection 
standard for children in child restraints but acknowledged that there 
are many unknowns. The agency sought comment on the suitability of the 
test procedures it was considering, on appropriate injury criteria for 
children in side impacts, on cost beneficial countermeasures, and on 
other issues. Additionally, after the ANPRM was published the agency 
evaluated possible mitigation concepts, such as adding padding material 
to the child restraint system. After reviewing the comments and the 
results of its post-ANPRM study, the agency has decided that the level 
and amount of effort needed to further develop and validate a side 
impact component for incorporation into FMVSS No. 213 far exceeds what 
could be accomplished within the time constraints of the TREAD Act. 
While an NPRM is not feasible at this time, NHTSA's research into 
improved side impact protection requirements for child restraints will 
continue as an ongoing agency program.
    The updates to the seat assembly are based on studies that NHTSA 
contracted to have done in response to the TREAD Act. This final rule 
makes the following changes: the seat bottom cushion angle is increased 
from 8 degrees off horizontal to 15 degrees; the seat back cushion 
angle is increased from 15 degrees off the vertical to 20 +/- 1 
degrees; the spacing between the anchors of the lap belt is increased 
from 222 millimeters (mm) to 400 mm in the center seating position and 
from 356 mm to 472 mm in the outboard seating positions; and the seat 
back of the seat assembly is changed, from a flexible seat back to one 
that is fixed, to represent a typical rear seat in a passenger car.
    The changes to the sled pulse are based on studies conducted in 
response to the TREAD Act. The test corridor is widened to make it 
easier for more test facilities to reproduce. The wider corridor 
extends the pulse from 80 milliseconds (ms) to approximately 90 ms in 
duration. The expanded corridor does not reduce the stringency of the 
test, and makes it easier to conduct compliance tests at speeds closer 
to 30 mph.
    This document enhances the use of test dummies in the evaluation of 
child restraints under Standard No. 213. NHTSA replaces most of the 
existing dummies with the new 12-month-old Child Restraint Air Bag 
Interaction (CRABI) dummy, and the state-of-the art Hybrid III 3- and 
6-year-old dummies. NHTSA also incorporates a weighted 6-year-old dummy 
(i.e., a Hybrid III 6-year-old dummy to which weights have been added) 
to test the structural integrity of child restraints recommended for 
use by children weighing 50 to 65 lb. Incorporation of the weighted, 62 
lb, dummy is viewed as an interim measure until such time as the Hybrid 
III 10-year-old dummy becomes available. Because the weighted dummy 
will be available for use in dynamic testing of child restraints for 
older children, this final rule extends the application of FMVSS No. 
213 to child restraint systems for children who weigh 65 lb or less.
    The agency has decided against adopting the scaled injury criteria 
developed in the context of the advanced air bag rulemaking of FMVSS 
No. 208. The agency was unable to confirm the existence of a safety 
problem that the scaled injury limits of FMVSS No. 208 would remedy. 
Relatedly, not enough is known about what modifications to child 
restraints could be made for the restraints to meet the proposed injury 
limits. In balancing the effects of meeting the scaled injury criteria 
against the possible impacts on the price of restraints, the agency 
determined that the scaled injury limits should not be added to FMVSS 
No. 213 at this time.
    NHTSA has examined the benefits and costs of these amendments, 
wishing to adopt only those amendments that contribute to improved 
safety, and mindful of the principles for regulatory decisionmaking set 
forth in Executive Order 12866, Regulatory Planning and Review. Its 
efforts to do so, however, have been limited by several factors. One is 
the limited time allowed by the schedule specified in the TREAD Act for 
initiating and completing this rulemaking. That has limited the amount 
and variety of information that the agency could obtain and testing 
that the agency could conduct to examine the efficacy of possible 
countermeasures under consideration and the effects of the various 
proposed amendments on child restraint performance. The other is the 
lack of specific accident data on children in motor vehicle crashes 
generally. For example, there is little available data on neck injury 
in children involved in motor vehicle crashes. Together, these 
limitations have made it difficult to assess and compare the benefits 
and costs of this rulemaking.
    The agency does not believe that updating the seat assembly and 
revising the crash pulse would affect dummy performance to an extent 
that benefits would accrue from such changes. The amendment of FMVSS 
No. 213 incorporating use of the new dummies in compliance tests, 
including testing with a weighted 6-year-old dummy, would result in a 
one-time cost of $1.68 million for manufacturers to purchase the new 
test dummies and $1.39 to $3.44 million to certify existing child 
restraints to the new dummies and test requirements. The annual long-
term costs are estimated to be $31,200 to test new models of booster 
seats (including built-in restraints) with a weighted 6-year-old dummy. 
We believe that use of the new dummies, in itself, would not 
necessitate redesign of child restraints.

II. Background

    Of the 31,910 passenger vehicle occupants killed in 2001, 1,003 
were children ages 0 through 10 years old. Four hundred ninety-seven 
(497) of these were less than 5 years old. The failure to use occupant 
restraints is a significant factor in most fatalities resulting from 
motor vehicle crashes for both adults and children. Of the 31,910 
passenger vehicle occupants killed in 2001, over half (55 percent) were 
unrestrained. Forty-six (46) percent of the 1,003 child occupant 
fatalities, ages 0 through 10 years old, were unrestrained. For child 
occupants less than 5 years old, 45 percent of the 497 fatalities were 
unrestrained.\2\ In 2001, 202 child occupants under 5 years of age were 
killed while restrained in child restraints, and another 32,000 were 
injured.
---------------------------------------------------------------------------

    \2\ Of the 2,787,000 passenger vehicle occupants injured in 
crashes in 2001, only 12 percent (324,000) were reported as 
unrestrained. The rates are about the same for child occupants. For 
children ages 0-10 years old, an estimated 147,000 were injured in 
motor vehicle traffic crashes in 2001, and 12 percent (18,000) of 
these children were unrestrained. Of the 59,000 child occupants less 
than 5 years of age who were injured, 11 percent (6,000) were 
unrestrained.
---------------------------------------------------------------------------

    NHTSA developed three strategies for reducing the number of 
children killed and injured in motor vehicle crashes in this country. 
(See Planning Document, 65 FR 70687; November 27, 2000; Docket NHTSA 
7938.) The first of these was a strategy designed to increase restraint 
use among all children and to ensure that the appropriate restraint 
systems are used correctly. The agency

[[Page 37622]]

estimated that if all children ages 0-4 years old were restrained in 
child restraint systems, 173 lives could have been saved in 1998. 
Additional studies have shown that as many as 68 additional deaths to 
children ages 0-6 years old could be prevented each year by eliminating 
misuse of child restraints. The agency conducts national campaigns to 
educate the public about the importance of buckling children into child 
restraint systems.
    The second strategy was to improve existing requirements for the 
performance and testing of child restraint systems. Since NHTSA first 
began regulating child safety seats in 1971, the agency has made 
numerous improvements to the original Federal safety standard. On a 
frequent basis, the agency has issued planning documents or has held 
public meetings on child passenger safety issues at the attention of 
the agency and the agency's long view of possible regulatory actions 
that might be taken in response. The public is invited to comment on 
the agency's plans. The November 2000 Planning Document announced that 
the agency planned to undertake rulemaking to update the bench seat and 
belt geometry used in Standard No. 213's compliance test, revise the 
crash pulse used in the test, incorporate state-of-the art infant, 3-
year-old and 6-year-old crash test dummies and child-specific injury 
criteria, and continue efforts working with the Society of Automotive 
Engineers in developing a 10-year-old child test dummy. The plan also 
stated that the agency would conduct research into possible side impact 
test requirements for child restraints and developing a test dummy 
appropriate for use in side impact tests. In addition, the plan 
announced that NHTSA would begin testing child restraints in full 
frontal and side impact vehicle crash tests under the agency's New Car 
Assessment Program.
    The third strategy called for improved mechanisms for getting 
safety information to consumers, to increase the likelihood that child 
restraints would be purchased and correctly used. The agency sought to 
improve the information it provided to consumers, both on the 
performance and proper use of child restraint systems, as well as on 
defect investigations and safety recalls.
    In November 2000, the TREAD Act was enacted. Section 14 of the 
TREAD Act directed NHTSA to initiate a rulemaking for the purpose of 
improving the safety of child restraints and included specific 
elements, listed below, that the agency had to consider as part of the 
rulemaking. Most of the elements for consideration had been included in 
NHTSA's Planning Document as part of the strategy for improving the 
safety of child restraints. Thus, Section 14 reaffirmed the importance 
of the agency's planned programs for amending Standard No. 213. 
Nonetheless, the TREAD Act had very tight deadlines for initiating and 
completing the rulemaking which also defined for the agency the actions 
it could take and complete within those deadlines.

III. The TREAD Act

    Section 14 of the TREAD Act directed NHTSA to initiate a rulemaking 
for the purpose of improving the safety of child restraints by November 
1, 2001, and to complete it by issuing a final rule or taking other 
action by November 1, 2002. The relevant provisions in Section 14 are 
as follows:
    (a) In General. Not later than 12 months after the date of 
enactment of this Act, the Secretary of Transportation shall initiate a 
rulemaking for the purpose of improving the safety of child restraints, 
including minimizing head injuries from side impact collisions.
    (b) Elements for Consideration. In the rulemaking required by 
subsection (a), the Secretary shall consider--
    (1) Whether to require more comprehensive tests for child 
restraints than the current Federal motor vehicle safety standards 
requires, including the use of dynamic tests that--
    (A) Replicate an array of crash conditions, such as side-impact 
crashes and rear-impact crashes; and
    (B) Reflect the designs of passenger motor vehicles as of the date 
of enactment of this Act;
    (2) Whether to require the use of anthropomorphic test devices 
that--
    (A) Represent a greater range of sizes of children including the 
need to require the use of an anthropomorphic test device that is 
representative of a ten-year-old child; and
    (B) Are Hybrid III anthropomorphic test devices;
    (3) Whether to require improved protection from head injuries in 
side-impact and rear-impact crashes;
    (4) How to provide consumer information on the physical 
compatibility of child restraints and vehicle seats on a model-by-model 
basis;
    (5) Whether to prescribe clearer and simpler labels and 
instructions required to be placed on child restraints;
    (6) Whether to amend Federal Motor Vehicle Safety Standard No. 213 
(49 CFR 571.213) to cover restraints for children weighing up to 80 
pounds;
    (7) Whether to establish booster seat performance and structural 
integrity requirements to be dynamically tested in 3-point lap and 
shoulder belts;
    (8) Whether to apply scaled injury criteria performance levels, 
including neck injury, developed for Federal Motor Vehicle Safety 
Standard No. 208 to child restraints and booster seats covered by in 
[sic] Federal Motor Vehicle Safety Standard No. 213; and
    (9) Whether to include [a] child restraint in each vehicle crash 
tested under the New Car Assessment Program.
    (c) Report to Congress. If the Secretary does not incorporate any 
element described in subsection (b) in the final rule, the Secretary 
shall explain, in a report to the Senate Committee on Commerce, 
Science, and Transportation and the House of Representatives Committee 
on Commerce submitted within 30 days after issuing the final rule, 
specifically why the Secretary did not incorporate any such element in 
the final rule.
    (d) Completion. Notwithstanding any other provision of law, the 
Secretary shall complete the rulemaking required by subsection (a) not 
later than 24 months after the date of the enactment of this Act.

IV. Responsible Regulation

    The agency developed its proposed and final rules responding to the 
TREAD Act while bearing in mind and in some cases, balancing, several 
compelling principles and considerations that generally come to the 
forefront in rulemaking in this area. These are discussed below.
    (a) When used, child restraints are highly effective in reducing 
the likelihood of death and or serious injury in motor vehicle crashes. 
NHTSA estimates (``Revised Estimates of Child Restraint 
Effectiveness,'' Hertz, 1996) that for children less than one-year-old, 
a child restraint can reduce the risk of fatality by 71 percent when 
used in a passenger car and by 58 percent when used in a pickup truck, 
van, or sport utility vehicle (light truck). Child restraint 
effectiveness for children between the ages 1 to 4 years old is 54 
percent in passenger cars and 59 percent in light trucks. The failure 
to use occupant restraints is a significant factor in most fatalities 
resulting from motor vehicle crashes. For child occupants less than 5 
years old, 45 percent of the 497 fatalities in 2001 were unrestrained.
    Over the past decade, the agency has sought to increase use of 
vehicle seat belt and child restraint systems. NHTSA conducts national 
campaigns to educate the public about the importance of buckling 
children into child restraint systems, supports efforts by state and

[[Page 37623]]

local organizations that wish to establish child safety seat fitting 
stations (locations within a community where parents and caregivers can 
learn how to install and properly use child restraints), and works with 
partners to train educators that can teach the public about using child 
restraints. If more child restraints were used, children's lives would 
certainly be saved in significant numbers.
    If child restraints were made more effective, some lives could also 
possibly be saved. However, in making regulatory decisions on possible 
enhancements, the agency must bear in mind the consumer acceptance of 
cost increases to an already highly-effective item of safety equipment. 
Any enhancement that would significantly raise the price of the 
restraints could potentially have an adverse effect on the sales of 
this voluntarily-purchased equipment. The net effect on safety could be 
negative if the effect of sales losses exceeds the benefit of the 
improved performance of the restraints that are purchased. Thus, to 
maximize the total safety benefits of its efforts to extend and upgrade 
its restraint requirements, the agency must balance those improvements 
against impacts on the price of restraints. The agency must also 
consider the effects of improved performance on the ease of using child 
restraints. If the use of child restraints becomes overly complex, the 
twin problems of misuse and nonuse of child restraints could be 
exacerbated.
    (b) Estimating the net effect on safety of this rulemaking, 
consistent with the principles for regulatory decisionmaking set forth 
in Executive Order 12866, Regulatory Planning and Review, was limited 
by several factors. One was the lack of specific accident data on 
children in motor vehicle crashes generally. Second, the limited time 
allowed by the schedule specified in the TREAD Act for initiating and 
completing this rulemaking limited the amount and variety of 
information that the agency could obtain and testing that the agency 
could conduct to examine the efficacy of possible countermeasures and 
the effects of various proposed amendments on child restraint 
performance. Together, these limitations made it difficult to assess 
and compare the benefits and costs of this rulemaking.
    (c) The rulemaking schedule imposed by the TREAD Act also limited 
the rulemaking to elements that could be completed within the statutory 
schedule. The development of an anthropomorphic test device 
representing a 10-year-old child could not be completed within the 
timeframe of the TREAD Act and so was not part of the rulemaking, 
notwithstanding its inclusion as an element for consideration in 
NHTSA's Planning Document and in Section 14 of the TREAD Act. 
Development of a seat cushion with different stiffness characteristics 
for the test seat assembly could not be completed and analyzed in time 
to be included in this rulemaking. Development of a side impact test 
procedure, injury criteria, and cost-effective countermeasures also 
could not be completed within the TREAD Act rulemaking schedule. Work 
is continuing in some of these areas. While ideally the agency would 
have wanted to address all related aspects of the standard, what could 
be accomplished in the near term was addressed and what could not but 
should will be pursued in the future.

V. Response to the TREAD Act

    Bearing in mind the principles and considerations discussed in the 
previous section, the agency initiated several actions following 
enactment of the TREAD Act. These are summarized below.

a. NPRM for This Final Rule

    On May 1, 2002 (67 FR 21806, docket 11707), the agency published a 
notice of proposed rulemaking (NPRM) proposing to incorporate five 
elements into the standard: (a) An updated bench seat used to 
dynamically test add-on child restraint systems; (b) a sled pulse that 
provides a wider test corridor; (c) improved child test dummies; (d) 
expanded applicability to child restraint systems recommended for use 
by children weighing up to 65 pounds; and (e) new or revised injury 
criteria to assess the dynamic performance of child restraints. The 60-
day comment period provided by the NPRM on the proposals was extended 
an additional 30 days in response to petitions from the Juvenile 
Products Manufacturers Association and ARCCA, Inc. 67 FR 44416; July 2, 
2002.
    The proposed updates to the seat assembly were based on studies 
that NHTSA contracted to have done in response to the TREAD Act. The 
NPRM proposed the following changes: the seat bottom cushion angle 
would be increased from 8 degrees off horizontal to 15 degrees; the 
seat back cushion angle would be increased from 15 degrees off the 
vertical to 22 degrees; the spacing between the anchors of the lap belt 
would be increased from 222 millimeters (mm) to 392 mm in the center 
seating position and from 356 mm to 472 mm in the outboard seating 
positions; and the seat back of the seat assembly would be changed, 
from a flexible seat back to one that is fixed, to represent a typical 
rear seat in a passenger car.
    The agency also proposed to widen the corridor of the sled pulse to 
make it easier for more test facilities to reproduce. The wider 
corridor extends the pulse from 80 milliseconds (ms) to approximately 
90 ms in duration. The agency believed that the expanded corridor would 
not reduce the stringency of the test, and would also make it easier to 
conduct compliance tests at speeds closer to 30 mph.
    The NPRM proposed two initiatives toward enhancing the use of test 
dummies in the evaluation of child restraints under Standard No. 213. 
NHTSA proposed to replace some of the existing dummies with the new 12-
month-old Child Restraint Air Bag Interaction (CRABI) dummy, and the 
state-of-the art Hybrid III 3- and 6-year-old dummies. NHTSA also 
proposed testing child restraints for older children with a weighted 6-
year-old dummy (i.e., a Hybrid III 6-year-old dummy to which weights 
have been added). The total weight of the dummy would be 62 lb. The 
agency sought to use the weighted dummy as an interim measure to test 
child restraints that are recommended for children weighing 50 to 65 
lb, until such time as a Hybrid III 10-year-old dummy now in 
development becomes available.
    The NPRM proposed to extend Standard No. 213 to apply to child 
restraint systems for children who weigh 65 lb or less. Restraints 
recommended for children weighing 50 to 65 lb would be tested with the 
weighted 6-year-old dummy.
    The proposal to use the new and scaled injury criteria of Standard 
No. 208 was based on research that the agency had done in support of 
the agency's May 2000 final rule on advanced air bag technology, which 
amended Standard No. 208 by, among other things, adjusting the criteria 
and performance limits to account for motor vehicle injury risks faced 
by different size occupants (65 FR 30680; May 12, 2000), as well as on 
NCAP and sled testing done in response to the TREAD Act. The NPRM 
proposed to adopt the scaled Head Injury Criterion (HIC) limits from 
the Standard No. 208 rulemaking into Standard No. 213, as well as the 
chest deflection and acceleration limits. The Nij neck criterion was 
also proposed to be added to Standard No. 213, but without the limits 
on axial force.
    NHTSA estimated that the proposal to use the new and scaled injury 
criteria of Standard No. 208 would prevent an estimated 3-5 fatalities 
and 5 MAIS 2-

[[Page 37624]]

5 non-fatal injuries for children ages 0-1 annually. In addition, the 
proposal would save 1 fatality and mitigate 1 MAIS 2-5 injury in the 4- 
to 6-year-old age group annually. The agency did not believe that 
updating the seat assembly and revising the crash pulse would affect 
dummy performance to an extent that benefits would accrue from such 
changes.
    NHTSA did not identify countermeasures to improve child restraint 
performance in frontal tests that would allow child restraint 
manufacturers to meet the proposed neck injury criterion. Consequently, 
the agency was unable to estimate the costs of such countermeasures. 
Comments were requested on possible countermeasures and their costs. 
The agency believed that the proposal to use new dummies in compliance 
tests, including testing with a weighted 6-year-old dummy, could result 
in increased testing costs for manufacturers that want to certify their 
restraints using the tests that NHTSA will use in compliance testing. 
NHTSA estimated that use of the new dummies and other changes to the 
test procedure would add testing costs of $2.72 million. The agency 
believed that those changes would not result in redesign of child 
restraints.

b. ANPRM On Side Impact Protection

    On May 1, 2002, concurrent with the publication of the NPRM and in 
further response to section 14(b) of the TREAD Act, NHTSA issued an 
advance notice of proposed rulemaking (ANPRM) seeking public comments 
on the agency's work on developing near-term a possible side impact 
protection standard for child restraint systems. 67 FR 21836, May 1, 
2002; Docket No. 02-12151.
    The ANPRM primarily addressed side impact protection for children 
in child restraints in the following areas: (a) Determination of child 
injury mechanisms in side impacts, and crash characteristics associated 
with serious and fatal injuries to children in child restraints; (b) 
development of test procedures, a suitable test dummy and appropriate 
injury criteria; and (c) identification of cost beneficial 
countermeasures. Uncertainties in these areas, together with the 
statutory schedule for this rulemaking, made it difficult for the 
agency to assess and make judgments on the benefits and costs of a 
rulemaking on side impact protection. The ANPRM also requested comments 
on the appropriateness of proposing to incorporate a rear impact test 
procedure into FMVSS No. 213 for rear-facing child restraint systems.
    The agency received approximately 17 comments on the ANPRM. 
Commenters expressed qualified support for NHTSA's efforts to enhance 
child passenger protection in side impact crashes, but were concerned 
about the uncertainties with respect to the three areas highlighted 
above. A number of commenters believed that a dynamic test should 
account for some degree of vehicle intrusion into the occupant 
compartment, which overall the tests that the agency had been 
considering did not.
    Following publication of the ANPRM, the agency began a program of 
child restraint systems side impact testing that continues today, for 
completion in fall 2003. Some of the side impact testing in which the 
agency is engaged is as follows:

--Initial evaluation of mitigation concepts, such as adding padding 
material to the child restraint system (CRS), modifying the size of the 
side wings of the CRS, effect of rigid lower anchorages and additional 
tethering of the CRS for rear-facing CRS in a side impact.
--Initial evaluation of mitigation concepts, such as adding padding 
material to the child restraint, modifying the size of the side wings 
of the CRS, rigid lower anchorages and additional tethering of the CRS 
for forward-facing CRS in a side impact.

    If the results from the above two evaluations are successful in 
reducing injury levels, NHTSA will consider conducting a test series to 
determine if the stiffer shoulder/upper arm area of the HIII 3-year-old 
dummy influences head/neck performance, as compared to the TNO Q3 dummy 
developed by a European test dummy manufacturer.
    Upon further consideration of the comments on the ANPRM and the 
agency's side impact test program, we have decided not to issue an NPRM 
and final rule on side and rear impact protection at this time and thus 
are withdrawing the action. A full explanation of the agency's reasons 
for this decision is set forth in a report to Congress that NHTSA has 
issued concurrently with today's final rule.\3\ To summarize, the 
agency found that for side crashes: (a) Data are not widely available 
as to how children are being injured and killed in side impacts (e.g., 
to what degree injuries are caused by intrusion of an impacting vehicle 
or other object); (b) potential countermeasures for side impact 
intrusion have not been developed; and (c) there is not a consensus on 
an appropriate child test dummy and associated injury criteria for side 
impact testing. There was widespread support for NHTSA to monitor the 
progress of the International Standards Organization (ISO) to develop a 
harmonized side impact test procedure. A preliminary draft of an ISO 
side impact test procedure includes specifications for an intruding 
door member. However, no dummies are available at the present time 
whose construction is designed for side impact validation. Given the 
lack of an approved test device, and corresponding injury criteria, a 
final version of an ISO test procedure is not expected in the near 
future.
---------------------------------------------------------------------------

    \3\ The report is issued in response to subsection 14(c) of the 
TREAD Act, which directs NHTSA to explain in a report to Congress 
why the agency did not incorporate any element described in 
subsection 14(b) in a final rule.
---------------------------------------------------------------------------

    The level and amount of effort needed to further develop and 
validate the ISO side impact test procedure far exceeds what could be 
accomplished within the time constraints of the TREAD Act. While an 
NPRM is not feasible at this time, NHTSA's research into improved side 
impact protection requirements for child restraints will continue as an 
ongoing agency program.

c. TREAD Programs on Labeling and Consumer Information

    Two other regulatory initiatives on child restraint systems were 
completed in response to Section 14 of the TREAD Act. Pursuant to Sec.  
14(b)(5) of the Act, the agency issued a final rule on October 1, 2002 
(67 FR 61523, Docket 10916) on Standard No. 213's labeling and owner's 
manual requirements. The final rule amends the format, location, and 
content of some of Standard No. 213's labeling requirements to make the 
labels and instructions clearer and simpler.
    In addition, pursuant to Sec.  14(g) of the Act, NHTSA published a 
final rule establishing an ease-of-use child restraint ratings program 
on November 2, 2002 (see 67 FR 67491; November 5, 2002, Docket 01-
10053). The ratings program constituted the first step toward enhancing 
the safety of children through a consumer information program. The 
program established no binding obligation on any manufacturer. Rather, 
it will inform consumers about the features of child restraints that 
make child restraints easier to use, and will evaluate each child 
restraint on those features.
    NHTSA is also continuing an evaluation of whether to establish two 
complementary consumer information programs. The first would be based 
on child restraint dynamic performance.

[[Page 37625]]

The second would involve expanding the agency's New Car Assessment 
Program to include consumer information on how vehicles do in 
protecting child occupants. The agency will be conducting two pilot 
programs in these areas to assess how the agency should proceed. These 
programs are described in detail in a Response to Comments, Notice of 
Final Decision accompanying the November 5, 2002 final rule (67 FR 
67448; Docket 01-10053-67).

VI. Post-NPRM Testing

    The agency conducted several research projects since publication of 
the NPRM in an effort to assess whether the proposed changes would 
reduce the safety currently afforded by child restraints. NHTSA 
conducted three test projects, which are fully discussed in sections 
VIII.a.1, VIII.c.1, and VIII.e.1 of this preamble. The first test 
project related to the effect the revisions to the test seat assembly 
might have on the dynamic performance of child restraints. Dummies 
currently specified in FMVSS No. 213 were tested with child restraints 
on the revised seat assembly, and the performance of the dummies was 
compared to that observed in compliance tests. The second test project 
related to assessing any performance differences that may exist between 
the Hybrid II and the Hybrid III dummies. The third test project 
involved evaluating whether child restraints tested with the Hybrid III 
dummies could meet the proposed scaled HIC, chest injury limits, and 
Nij measures. Reports relating to these projects have been placed in 
the docket for this rulemaking.

VII. Summary of Comments on the NPRM

    NHTSA received approximately 35 comments on the May 1, 2002 NPRM 
for this final rule. Commenters included child restraint manufacturers, 
motor vehicle manufacturers, motor vehicle dealers and other industry 
associations, child passenger safety consumer groups, the National 
Transportation Safety Board, child safety research and testing 
organizations, and private individuals. The Juvenile Products 
Manufacturers Association (JPMA) conducted a series of 80 sled tests of 
child restraints in response to the NPRM and included the results of 
the testing in its comment.

a. General Comments on the Proposals

    The commenters generally expressed support for the regulatory goals 
of the NPRM to enhance child passenger safety.\4\ However, many 
underscored concerns that the rulemaking undertaken by the agency in 
response to the TREAD Act could possibly indirectly cause a reduction 
in child passenger safety, particularly with respect to applying new 
(neck loading, chest deflection) and revised (scaled HIC and chest 
acceleration limits) injury criteria used to assess the dynamic 
performance of child restraints. Commenters expressing these concerns 
were diverse. The JPMA, representing child restraint manufacturers 
Britax, Cosco, Evenflo, Graco/Century, and Peg Perego, believed that 
NHTSA should be concerned about ``unintended consequences of multiple, 
unevaluated changes to 213, as well as the adverse consequences of 
substantial cost increases.'' In a separate comment on the NPRM, 
Evenflo expressed opposition to ``revisions that do not have proven 
likelihood of enhancing child passenger safety on an aggregate basis.'' 
Evenflo urged: ``Our goal should be to adopt changes that will 
definitively enhance child passenger safety, not to undertake changes 
solely for the purpose of making changes.'' Graco was concerned that 
some portions of the proposed revisions may have little benefit to 
child passenger safety and may ``negatively affect the past efforts of 
the agency.'' The commenter said it assumed that costs of child 
restraints will increase ``because of increased testing costs and most 
likely increased parts or the use of more advanced technology that will 
enable the restraints to meet new requirements.'' The commenter was 
concerned that, as the child restraint costs rise, the rates of child 
restraint use may fall.
---------------------------------------------------------------------------

    \4\ Several commenters believed that the NPRM did not ``meet the 
spirit intended by Congress'' in enacting Section 14 of the TREAD 
Act because the estimated benefits of the proposed changes were at 
most 6 fatalities and 6 serious injuries annually (quote from 
ARCCA's comment, page 2.) Stephen Syson (Syson-Hille & Associates), 
Martha Bidez (Bidez & Associates) and ARCCA suggested that the 
agency undertake rulemaking beyond the proposals of the NPRM. Among 
the suggested rulemaking were the following from these commenters: 
the prohibition of lap belts; require manufacturers to put child-
safe restraints in cars; recall all low-shield booster seats; 
require that Standard No. 208 (49 CFR Sec.  571.208) criteria for 
children be met in every passenger seating position; require 
manufacturers to label vehicle seats that do not meet Standard No. 
213 requirements without a child restraint in place; require that 
``survival space'' be maintained in the rear seat in rear impact 
crashes; and require all seats, seat belts and child restraints be 
designed to prevent submarining and to retain occupants under all 
collision circumstances; require vehicles to provide a minimum 
allowable clearance for all seating positions where a child 
restraint system can be installed; require child restraints to 
provide both upper and lower body restraint on the hard boney 
portions of the body; amend Standard No. 213 to limit protrusions 
and sharp corners contactable in any crash and to improve padding on 
back and side wings.
    The rulemakings suggested by these commenters go beyond those 
included in the NPRM for consideration by the agency within the 
tight deadlines of the act. The suggestions will be considered 
suggestions for future rulemaking. Copies of the comments will be 
placed in Docket 7938 (NHTSA plan for future work on child passenger 
safety).
---------------------------------------------------------------------------

    This concern was echoed by other commenters. TraumaLink at the 
Children's Hospital of Philadelphia stated that data collected through 
its ``Partners for Child Passenger Safety'' study indicate that 
children in child restraints do extremely well in all types of crashes. 
``The extremely low injury rate in child restraints indicates that 
despite substantial misuse, these devices perform exceedingly well 
across the range of crash severities and directions of impact * * * It 
is important to consider the unintended consequences of these 
[proposed] changes, both in terms of inadvertent reduction in the 
current excellent performance of the CRS [child restraint system] or 
the resultant increase in cost.'' The National Safe Kids Campaign (Safe 
Kids) urged the agency ``to be mindful of the practicalities and costs 
associated with changes that might overly burden child restraint and 
vehicle manufacturers, thereby requiring them to discontinue certain 
product models or pass on unreasonable costs to consumers. Child safety 
seats must remain both affordable and safe.'' The American Academy of 
Pediatrics stated, ``While the Academy strongly supports the proposed 
measures to make child restraints even more protective than they are 
today, these improvements cannot come at the expense of fewer children 
using child restraints or more children using outdated car safety 
seats.'' The Association of International Automobile Manufacturers, 
Inc., expressed concern that ``the lack of use or the misuse of child 
restraint systems presents a far greater opportunity to improve child 
passenger safety than seeking enhanced performance of child restraint 
systems, particularly if the consequence of the enhanced performance is 
decreased use and increased misuse.''
    Safe Kids also expressed concern that increased prices of child 
restraints could affect State child restraint use laws. Safe Kids 
stated that most parents and caregivers will be expected to purchase a 
minimum of two or three restraints for each child to comply with 
evolving State child restraint use laws that extend coverage to more 
and more children. ``As those seats become more expensive, legislators 
may be reluctant to make important legislative

[[Page 37626]]

improvements to their states child restraint laws.''

b. Updated Bench Seat

    There was unanimous support for amending Standard No. 213's 
specifications for the test seat assembly used to test child restraints 
in the agency's compliance tests. Almost all of the commenters believed 
that the test seat assembly should be more representative of the seats 
of newer passenger vehicles. Two commenters (Martha Bidez, Public 
Citizen) had an opposing view. Ms. Bidez believed that the seat 
assembly should either have features representing seats in the average 
age vehicle in the U.S. (which the commenter stated is 9 years old) or 
features that present the most demanding (``worst case'') conditions 
under which child restraints should be tested. Public Citizen suggested 
that the agency should adjust its testing, or create another test, that 
will measure the effectiveness of child restraints in older cars.
    Amending the seat cushion angle by increasing it from 8 degrees off 
horizontal to 15 degrees was generally supported, as was amending the 
seat back angle by increasing it from 15 degrees off vertical to 22 
degrees. Several commenters viewed these changes as aligning the bench 
seat more with the ECE Regulation 44 seat assembly bench and suggested 
that the agency completely use the ECE Regulation 44 seat dimensions.
    Most commenters agreed with the proposals for amending the seat 
belts on the test seat assembly. Some commenters expressed concerns 
about certain aspects of the test seat assembly's seat belts that were 
not addressed by the NPRM, such as the vertical location of the lap 
belt anchorages.
    On the other hand, commenters did not see eye-to-eye on the 
proposal to change the seat back to represent a fixed vehicle seat. 
Supporters of the change believed that a fixed seat back replicates 
today's seat back in passenger cars and harmonizes with the test bench 
setups for ECE, Canadian and Australian regulations. Some commenters 
were concerned that not enough was known about how fixing the seat back 
would affect child restraint system performance, while others opposed 
the proposal believing that fixing the seat back would result in a less 
rigorous test condition.
    Several commenters responded to the NPRM's request for comments on 
the agency's decisions against changing, at this time the length of the 
bench seat's bottom seat cushion, including a floor to the seat 
assembly; and changing the stiffness of the seat assembly's cushion. A 
number of commenters believed that the stiffness of the seat cushion 
has a strong effect on child restraint performance. Some commenters 
were uncertain whether performance would be affected and suggested that 
testing and research be completed before changing the foam.

c. Crash Pulse

    The comments focused generally on the issues of the sled pulse 
shape (widening of the corridor) and severity. Many commenters agreed 
with the agency that widening the corridor of the sled pulse from 80 
milliseconds (ms) to approximately 90 ms in duration would allow more 
laboratories to run the compliance test without decreasing the 
effectiveness of the testing. However, child restraint manufacturers 
expressed concern that widening the corridor will make the standard 
more stringent, because child restraint manufacturers will have to 
design products that can comply at the new extremes of the compliance 
corridor. These commenters also believed that a wider test corridor 
will necessarily lead to more lab-to-lab variability during 
certification and compliance testing, which, the commenters stated, 
increases the compliance burden on manufacturers.
    ARCCA, Inc. believed that the Standard No. 213 pulse is actually 
less severe than all of the 30 mph barrier test pulses from actual 
vehicles, and that the standard's pulse severity should be increased. 
All other commenters did not want to increase the severity of the crash 
pulse. Many expressed the view that the velocity change should not be 
raised because the current test is already reflective of the top few 
percent of crashes. A number of commenters believed that the crash 
pulse should be reduced in severity because the most frequent crashes 
involving children in child restraints are those with lower crash pulse 
severities than the test pulse. Others believed that a relatively 
severe, ``worst case'' scenario should be replicated.

d. New Dummies

    Commenters generally supported using the CRABI and Hybrid III 3-
year-old dummies in Standard No. 213 compliance tests, in place of the 
9-month and Hybrid II 3-year-old dummies now used by the agency. 
However, a number of commenters expressed concerns that the Hybrid III 
6-year-old dummy's neck was too flexible for use in testing child 
restraints. These commenters suggested that the agency continue its use 
of the Hybrid II 6-year-old dummy rather than use the Hybrid III dummy 
in its place. Most commenters objected to using the weighted Hybrid III 
6-year-old dummy (weighted to 62 lb) to assess injury reference values 
in compliance tests of child restraints recommended for use by children 
weighing over 50 lb. Most believed that the dummy's weighting produced 
a dummy that was unrepresentative of a 62 lb child. Some were concerned 
that the weights could interfere with the proper functioning of the 
dummy's instrumentation. Some of these commenters suggested that the 
dummy should be used only to assess the structural integrity of child 
restraints in the standard's dynamic test, and not the capability of 
the restraint to limit head excursion or forces to the dummy's head, 
neck or chest areas.

e. Application of the Standard

    Of the commenters addressing application of the standard, a 
majority supported increasing the weight limit contained in the ``child 
restraint system'' definition. Most of these commenters supported 
increasing the weight limit to 65 lb with a future increase to 80 lb 
upon introduction of the 10-year-old dummy. A few commenters opposed 
establishing 65 lb as an intermediate step in favor of amending the 
standard directly to 80 lb. There were also a few divergent comments on 
whether the agency should extend the regulation to a maximum weight 
beyond that of the heaviest dummy used in the standard.

f. Injury Criteria

    The agency received widely divergent comments on the proposal to 
limit measurement of HIC to 15 milliseconds and to use the injury 
criteria of Standard No. 208 that were scaled for children. The 
Alliance, UMTRI and SafetyBeltSafe supported the use of a 15 ms limit 
on the head injury criterion (HIC) limit as a more realistic way to 
assess head and brain injury, with the lower HIC values proposed for 
each dummy. JPMA stated that it was willing to consider supporting a 15 
ms limit (HIC 15), if the agency can undertake research to assure that 
there will not be unintended consequences from countermeasures needed 
to meet HIC 15. However, JPMA did not support the other proposed new 
injury criteria, including the scaled HIC values. The commenter stated 
that the tests of child restraints it conducted with the proposed CRABI 
and Hybrid III dummies produced injury reference values that exceeded 
the proposed limits, which the commenter said is a concern given the 
high level of

[[Page 37627]]

effectiveness of current child restraints. The commenter suggested that 
it might be more feasible to use the FMVSS No. 208 criteria in FMVSS 
No. 213 if the agency were to specify a ``more realistic crash pulse 
for FMVSS No. 213, such as the one contained in the FMVSS No. 208 sled 
test.'' Graco was concerned that some seats that have historically 
performed well in the real world and in compliance testing would fail 
the new criteria.
    A few commenters supported while others opposed the proposals to 
adopt a new chest deflection criterion and to adopt the chest 
acceleration limits that were scaled for children and incorporated into 
FMVSS No. 208. JPMA, TraumaLink, UMTRI, SafetyBeltSafe and others 
opposed incorporation of the proposed chest deflection and reduced 
chest acceleration limits, because these types of injuries do not occur 
in children in child restraint systems. These commenters and others 
suggested that the agency collect data on chest deflection to establish 
a database that could be used to evaluate these measures more in the 
future.
    Virtually all parties commenting on this aspect of the proposal 
opposed the modified Nij neck criterion (modified from the criterion in 
FMVSS No. 208 in that the limits on axial force were excluded). JPMA, 
SafetyBeltSafe, UMTRI, TraumaLink and others did not support adopting 
the proposed Nij criterion at this time because the relationship 
between the criterion and real-world injuries ``under the type of 
loading simulated by FMVSS 213'' is ``not well established'' (quoting 
UMTRI). SafetyBeltSafe believed that neither Nij as proposed nor Nij 
with a limit on tension should be used as a compliance criterion unless 
these are proven to be useful predictors of child neck injury. The 
Insurance Institute for Highway Safety (IIHS) was concerned that 
studies of real-world crashes indicate that neck injuries due to 
inertial forces appear to be rare, and, the commenter stated, it is not 
clear how child restraints could be better designed to lower neck 
injury measures.

VIII. Amendments

a. Updated Bench Seat

1. Post-NPRM Test Program
    As discussed in the NPRM, NHTSA had initiated a test program in 
response to the TREAD Act to assess seat parameters of production 
seats, working with the U.S. Naval Air Warfare Center Aircraft Division 
at Patuxent River, Maryland (PAX). PAX analyzed seat geometry data, 
including seat cushion angle, seat back angle, seat cushion length, 
seat back length, tether anchor locations, child restraint anchorage 
system anchor locations, and seat belt locations.
    After publication of the NPRM, PAX conducted a series of dynamic 
tests using a revised test seat assembly that had been constructed 
incorporating the changes to the test seat assembly proposed in the 
NPRM. These tests were conducted with the dummies currently specified 
in FMVSS No. 213 (the newborn and TNO 9-month, and Hybrid II 3- and 6-
year-old dummies), and with various types of child restraints (rear-
facing infant only, rear- and forward-facing convertible, forward-
facing ``hybrid boosters'' (a child restraint that can be used as a 
forward-facing restraint with harness for toddlers up to 40 lb and as a 
belt-positioning booster with children over 40 lb), and both backless 
and high-back boosters).
    The results from this series of dynamic sled tests were compared to 
actual compliance tests that the agency had conducted to determine what 
effect, if any, the revisions to the test seat assembly might have on 
the dynamic performance of child restraints. NHTSA compared 
measurements taken for seat back rotation in rear-facing tests, and 
HIC, chest acceleration, and head and knee excursion in forward-facing 
tests. All of the proposed changes were simultaneously incorporated 
into the test seat assembly, and were not individually assessed for its 
effect on child restraint performance.
    i. Seat Back Rotation. The effect of the revised test seat assembly 
on measured seat back rotation in rear-facing tests did not show a 
clear pattern.
    Rear-facing tests were conducted using the revised test seat 
assembly with rear-facing infant only seats using the newborn dummy, 
and rear-facing convertible restraints using the newborn and Hybrid II 
9-month-old dummies. In tests of rear-facing restraints, HIC and chest 
acceleration are not currently measured, since the newborn and 9-month-
old dummies are not instrumented. Further, head and knee excursion are 
not measured. The only measured parameter in testing rear-facing child 
restraints is provided in S5.1.4 of FMVSS No. 213, which specifies that 
when a rear-facing child restraint is tested, the angle between the 
system's back support surface for the child and the vertical shall not 
exceed 70[deg].
    The seat back rotation measured in these tests is compared to the 
seat back rotation measured in NHTSA compliance tests of the identical 
child restraints in Table 1 below.

                                Table 1.--Seat Back Rotation in Rear-Facing Tests
----------------------------------------------------------------------------------------------------------------
                                                                         Seat back rotation (degrees)
                                                                         relative to vertical-- Test
         Child restraint                  Type              Dummy               seat assembly            Change
                                                                      ---------------------------------    (%)
                                                                            Existing         Revised
----------------------------------------------------------------------------------------------------------------
Evenflo On-My-Way...............  Infant only.........  Newborn......  43................         51.5     +19.8
Century 560.....................  Infant only.........  Newborn......  46................         42.5      -7.6
Evenflo On-My-Way...............  Infant only.........  9-month......  57................         53.9      -5.4
Century 560.....................  Infant only.........  9-month......  52................         52.9      +1.7
Century Accel...................  Convertible.........  Newborn......  Not tested........         50.7       N/A
Century STE 2000................  Convertible.........  Newborn......  Not tested........           40       N/A
Cosco Triad (LATCH \5\ )........  Convertible.........  Newborn......  Not tested........         43.1       N/A
Century STE 2000................  Convertible.........  9-month......  42................         50.6     +20.5
Cosco Touriva...................  Convertible.........  9-month......  51................           63    +23.5
----------------------------------------------------------------------------------------------------------------
\5\ ``LATCH'' stands for ``Lower Anchors and Tethers for Children,'' a term that was developed by manufacturers
  and retailers to refer to the standardized child restraint anchorage system required by FMVSS No. 225. This
  preamble uses the term to describe either an FMVSS No. 225 anchorage system in a vehicle or a child seat that
  attaches to an FMVSS No. 225 child restraint anchorage system.


[[Page 37628]]

    The data indicated no clear effect of the revised test seat 
assembly on measured seat back rotation in rear-facing tests. In tests 
using the newborn dummy and two different rear-facing infant-only child 
restraints, the seat back rotation angle increased by 19.8 percent over 
that measured in the comparable compliance test in one, and decreased 
by 7.6 percent in the other. When the same infant-only seats were 
tested rear facing with the 9-month-old dummy, the restraint that had 
previously shown increased seat back rotation with the newborn dummy 
decreased by 5.4 percent over that measured in the comparable 
compliance test, while the restraint that had shown decreased seat back 
rotation with the newborn dummy increased by 1.7 percent over that 
measured in the comparable compliance test. In all cases, the measured 
seat back rotation was well under the FMVSS No. 213 limit of 70[deg].
    Tests were conducted using the revised test seat assembly on three 
different rear-facing convertible child restraints with the newborn 
dummy. In each case, the measured seat back rotation angle was well 
below the FMVSS No. 213 limit.
    PAX also conducted tests of two different rear-facing convertible 
child restraints with the 9-month-old dummy using the revised test seat 
assembly. In each of these tests, the seat back rotation increased by 
at least 20 percent over that measured in the comparable FMVSS No. 213 
compliance tests conducted on the existing test seat assembly. Again, 
however, the rotation was within the allowable limits.
    ii. HIC Measurements. Generally speaking, HIC increased in tests 
with the Hybrid II 3-year-old dummy, and decreased in tests with the 6-
year-old.
    Sled tests were conducted using the revised test seat assembly with 
the Hybrid II 3-year-old dummy in forward-facing convertible 
restraints, and in forward-facing hybrid boosters using the restraint's 
internal harness (in the toddler seat mode), and with the Hybrid II 6-
year-old dummy in both backless and high back belt-positioning booster 
restraints. The HIC measured in these tests is compared to the HIC 
measured in NHTSA compliance tests of the same model child restraints 
in Table 2 below.

                            Table 2.--HIC in Tests of Forward-Facing Child Restraints
----------------------------------------------------------------------------------------------------------------
                                                                             HIC[infin] Test seat
                                                          Dummy  (Hybrid           assembly
         Child restraint                   Type                 II)       --------------------------  Change (%)
                                                                             Existing     Revised
----------------------------------------------------------------------------------------------------------------
Cosco Touriva....................  Convertible.........  3-year..........          500          703        +40.6
Century Accel....................  Convertible.........  3-year..........          480          627        +30.5
Century Breverra.................  Hybrid Booster......  3-year..........          659          670         +1.6
Cosco High Back Booster..........  Hybrid Booster......  3-year..........          535          446        -16.6
Cosco Grand Explorer.............  Backless BPB........  6-year..........          438          267        -39.0
Cosco Grand Explorer.............  Backless BPB........  6-year..........          438          328        -25.1
Century Breverra.................  High-back BPB.......  6-year..........          308          209        -32.0
Cosco High Back Booster..........  High-back BPB.......  6-year..........          399          381         -4.6
----------------------------------------------------------------------------------------------------------------

    The effect of the revised seat assembly on HIC measurements appear 
to be varied, and largely dependent on the dummy used in the testing. 
In three of four tests conducted with the 3-year-old dummy, the 
measured HIC was higher using the revised test seat assembly as 
compared to compliance tests performed on the existing test seat 
assembly. This includes both tests conducted using forward-facing 
convertible restraints, and one of two tests using a forward-facing 
hybrid booster with its internal harness system. However, in each of 
four tests conducted with the 6-year-old dummy, two each with backless 
boosters and high back boosters, the measured HIC was lower than in the 
identical compliance tests conducted on the existing test seat 
assembly. Overall, some measurements differed by as much as +/- 40 
percent between tests conducted on the two different test seat 
assemblies. All HIC measurements were well within the existing limit of 
1000.
    iii. Chest Acceleration. Chest acceleration measurements were 
recorded using the Hybrid II 3- and 6-year-old dummies in the same 
series of tests outlined in Table 2 above. Table 3 details the recorded 
chest acceleration in these tests as well as the comparable compliance 
tests of the identical child restraints. The measured chest 
accelerations decreased in each of the tests using the 3-year-old dummy 
in the revised test seat assembly. The measured chest accelerations 
generally increased in tests using the 6-year-old dummy in the revised 
test seat assembly.

                    Table 3.--Chest Acceleration in Tests of Forward-Facing Child Restraints
----------------------------------------------------------------------------------------------------------------
                                                                           Chest acceleration (g)--
                                                           Dummy (Hybrid       Test seat assembly
         Child restraint                   Type                 II)       -------------------------- Change  (%)
                                                                             Existing     Revised
----------------------------------------------------------------------------------------------------------------
Cosco Touriva....................  Convertible.........  3-year-old......           42         40.4         -3.8
Century Accel....................  Convertible.........  3-year-old......           46         26.8        -41.7
Century Breverra.................  Hybrid Booster......  3-year-old......           40         29.2        -27.0
Cosco High Back Booster..........  Hybrid Booster......  3-year-old......           44         41.6         -5.5
Cosco Grand Explorer.............  Backless BPB........  6-year-old......           44         49.2        +11.8
Cosco Grand Explorer.............  Backless BPB........  6-year-old......           44         38.6        -12.3
Century Breverra.................  High-back BPB.......  6-year-old......           33         35.1         +6.4
Cosco High Back Booster..........  High-back BPB.......  6-year-old......           40         42.4         +5.5
----------------------------------------------------------------------------------------------------------------


[[Page 37629]]

    All chest acceleration measurements recorded were well within the 
current limit of 60 g's maximum. It is noted, however, that while most 
chest acceleration measurements were comparable in magnitude between 
the two test seat assemblies, there was one test in which the measured 
values differed by 42 percent for the same child restraint. iv. Head 
Excursion. It is not evident whether use of the revised test seat 
assembly will have a positive or negative effect on measured head 
excursion.
    In the tests outlined in Tables 2 and 3, supra, head excursion was 
measured. In addition, head excursion was measured in sled tests 
performed with the TNO 9-month-old dummy on two different forward-
facing convertible restraints. Head excursion was compared to the head 
excursion measured in compliance tests of the identical child 
restraints using the same dummies. Table 4 provides this comparison.

                      Table 4.--Head Excursion in Tests of Forward-Facing Child Restraints
----------------------------------------------------------------------------------------------------------------
                                                                             Head excursion (mm)--
                                                                              Test seat assembly
         Child restraint                  Type         Dummy  (Hybrid II) -------------------------- Change  (%)
                                                                             Existing     Revised
----------------------------------------------------------------------------------------------------------------
Cosco Touriva...................  Convertible........  9-month-old.......          432          434         +0.6
Century Accel...................  Convertible........  9-month-old.......          483          396        -17.9
Cosco Touriva...................  Convertible........  3-year-old........          660          498        -24.6
Century Accel...................  Convertible........  3-year-old........          635          495        -22.0
Century Breverra................  Hybrid Booster.....  3-year-old........          483          572        +18.4
Cosco High Back Booster.........  Hybrid Booster.....  3-year-old........          432          572        +32.4
Cosco Grand Explorer............  Backless Booster...  6-year-old........          381          363         -4.7
Cosco Grand Explorer............  Backless Booster...  6-year-old........          381          457        +20.0
Century Breverra................  High-back Booster..  6-year-old........          457          500         +9.4
Cosco High Back Booster.........  High-back Booster..  6-year-old........          432          447         +3.5
----------------------------------------------------------------------------------------------------------------

    In three of four tests conducted using forward-facing convertible 
child restraints, a decrease in head excursion was observed in tests 
using the revised test seat assembly. However, in tests conducted on 
the revised seat assembly using forward-facing hybrid boosters, 
backless and high back belt-positioning booster seats, a marginal 
increase in head excursion was observed. All measured head excursions, 
on the existing and revised test seat assemblies, were well within the 
established 813 mm limit prescribed in FMVSS No. 213.\6\
---------------------------------------------------------------------------

    \6\ Excursions are measured from Point Z identified in Figure 1B 
of FMVSS No. 213, which is located in the same place on both the 
existing and revised test seat assemblies.
---------------------------------------------------------------------------

    v. Knee Excursion. For the tests of forward-facing child restraints 
outlined in Table 4 above, NHTSA also measured the dummy's knee 
excursion. These results were compared to the knee excursion measured 
in compliance tests of the identical child restraints using the same 
dummies. The knee excursion measurements did not demonstrate a direct 
correlation between tests conducted with the revised test seat assembly 
versus the existing test seat assembly, or with the type of child 
restraint used or the test dummy used. Table 5 presents the results. As 
with the other injury criteria discussed above, all knee excursion 
measurements were well within the established 915 mm limit prescribed 
in FMVSS No. 213.

                      Table 5.--Knee Excursion in Tests of Forward-Facing Child Restraints
----------------------------------------------------------------------------------------------------------------
                                                                             Knee excursion (mm)--
                                                                              Test seat assembly
         Child restraint                  Type         Dummy  (Hybrid II) -------------------------- Change  (%)
                                                                             Existing     Revised
----------------------------------------------------------------------------------------------------------------
Cosco Touriva...................  Convertible........  9-month-old.......          483          546        +13.2
Century Accel...................  Convertible........  9-month-old.......          559          485        -13.2
Cosco Touriva...................  Convertible........  3-year-old........          813          671        -17.5
Century Accel...................  Convertible........  3-year-old........          762          681        -10.7
Century Breverra................  Hybrid Booster.....  3-year-old........          584          696        +19.1
Cosco High Back Booster.........  Hybrid Booster.....  3-year-old........          635          660         +4.0
Cosco Grand Explorer............  Backless Booster...  6-year-old........          686          610        -11.1
Cosco Grand Explorer............  Backless Booster...  6-year-old........          686          653         -4.8
Century Breverra................  High-back Booster..  6-year-old........          610          500        -17.9
Cosco High Back Booster.........  High-back Booster..  6-year-old........          686          701         +2.2
----------------------------------------------------------------------------------------------------------------

    vi. Summary of PAX Testing. Overall, while differences were seen in 
tests using identical child restraints on the existing versus the 
revised test seat assembly, NHTSA did not identify any specific trends 
along specified parameters, i.e., child restraint type, dummy, etc. All 
of the measured injury criteria in the tests were well within the 
established limits of FMVSS No. 213. This leads the agency to conclude 
that the changes to the standard test seat assembly will not have a 
significant effect on compliance test results of child restraint 
systems that meet the current requirements of the standard. 
Manufacturers will not need to redesign their restraints due to the 
changes in the seat assembly.
2. Response to Comments
    There was unanimous support for amending Standard No. 213's 
specifications for the test seat assembly

[[Page 37630]]

used to test child restraints in the agency's compliance tests. Almost 
all of the commenters believed that the test seat assembly should be 
more representative of the seats of newer passenger vehicles.
    i. Seat Back and Cushion Angles. Amending the seat cushion angle by 
increasing it from 8 degrees off horizontal to 15 degrees was generally 
supported. Several commenters viewed these changes as aligning the 
bench seat more with the ECE Regulation 44 seat assembly bench. Ford 
believed that the proposed change to the seat cushion angle would help 
make rigid attachment LATCH infant seats commercially viable in the 
U.S., and would help facilitate the use of infant restraints by 
reducing the need for consumers to add towels or pool noodles as 
spacers under the restraints. Ms. Bidez and Public Citizen opposed the 
proposed change to the seat cushion angle, stating that seat cushion 
angle should represent the average angle of a 9-year-old vehicle, not a 
new vehicle. Ms. Bidez stated that older seat cushions are more 
horizontal and do not contain any anti-submarining structural 
components.
    The agency has decided to revise the seat cushion angle as 
proposed. Increasing seat cushion angle from 8 degrees off horizontal 
to 15 degrees will make the seat assembly more representative of 
currently manufactured vehicle seats and will reduce or eliminate the 
need for supplementary devices, such as rolled towels or swimming 
noodles, now being used with infant seats to compensate for the 
difference in seat cushion angle of the current seat assembly and new 
vehicle seats. The agency does not agree with Ms. Bidez and Public 
Citizen that the seat assembly should be representative of seats in 9-
year-old vehicles. Such a rearward-looking approach ensures the 
obsolescence of the standard, since seats in the vehicle fleet are 
already in the process of being replaced by the seats of more modern 
design.
    UMTRI expressed concern that tests of child restraints on a seat 
assembly with a seat cushion at the proposed 15 degree angle to 
horizontal generally resulted in decreased head excursion values of 
about two inches and increased chest accelerations by an average of 4 
g. UMTRI suggested reducing the allowable head excursion limit in 
Standard No. 213 by two inches to compensate for the change. JPMA 
disagreed with UMTRI's comment that the head excursion limit should be 
changed, stating its belief that there is no difference in safety since 
the reference point from which head excursion is measured is unchanged. 
JPMA further stated that--

the fact that the increased angle allows the child's head to travel 
a longer distance in the real world will permit the manufacturers to 
utilize that additional movement to manage some of the crash energy 
without making other, perhaps less desirable, changes to other 
restraint parameters. For example, the harness system could include 
measures and/or devices to add energy absorption similar to vehicle 
retractor torsional load limiters, which were implemented with air 
bags as a means to reduce chest compression. Such devices require 
that a small amount of additional head excursion be permitted in the 
real world to achieve a longer ride-down and take advantage of the 
vehicle's `crumple zone.' * * *

    The agency does not agree that testing on the new seat assembly 
will result in across-the-board reductions in dummy head excursions as 
compared to head excursions of dummies tested on the current assembly. 
It is not evident from the agency's test data that use of the revised 
test seat assembly will have a positive or negative effect on measured 
head excursion. Table 4, supra, provides test results comparing head 
excursion measurements in a total of 10 tests using the revised test 
seat assembly and using the existing test seat assembly (compliance 
test results). These tests were conducted using (1) the 9-month-old 
dummy in two different forward-facing convertible restraints, (2) the 
3-year-old dummy in two forward-facing convertible restraints and two 
forward-facing hybrid booster restraints, and (3) the 6-year-old dummy 
in two backless boosters and in two high back belt-positioning 
boosters. In three of four tests conducted using forward-facing 
convertible child restraints, a decrease in head excursion was observed 
in tests using the revised test seat assembly. However, in tests 
conducted on the revised seat assembly using forward-facing hybrid 
boosters, backless and high back belt-positioning booster seats, a 
marginal increase in head excursion was observed.
    While differences of up to +32.4 percent and -24.6 percent were 
measured in tests using the revised and existing test seat assemblies, 
there was no distinctive trend across dummy or child restraint types. 
Thus, the agency cannot conclude that the new seat assembly necessarily 
results in a less rigorous test of a child restraint's ability to limit 
head excursion as compared to the existing seat assembly. Further, all 
measured head excursions on the existing and revised test seat 
assemblies in NHTSA's program were well within the established 813 mm 
limit prescribed in FMVSS No. 213. Thus, the agency does not believe 
that there has been a showing of a safety need to reduce the head 
excursion limit to take account of the effect of testing on the new 
test assembly.
    In response to JPMA's comment about increased head excursion 
benefiting overall child restraint performance due to increased ``ride 
down'' of crash forces, the agency agrees that generally speaking, 
increased ride down can help reduce head, neck and chest accelerations. 
However, increased ride down obviously must not come at the cost of 
increased risk of head impacts due to excessive head excursions in a 
crash. Thus, the agency does not concur with any implication that head 
excursions beyond what is permitted by Standard No. 213 is acceptable. 
The agency is concerned that child restraints that might meet the head 
excursion requirements of the standard when tested on the new test seat 
assembly might allow excessive head excursion when used in actual 
vehicles whose seat cushions are more like the current seat assembly. 
The agency asks the public for help in monitoring this situation and 
providing information of a real world problem should one occur. If 
there are unreasonable excessive head excursions due to child 
restraints being used on vehicle seats that are flatter than the 
revised seat assembly, reducing the head excursion limit of the 
standard will be considered by the agency.
    Amending the seat back angle by increasing it from 15[deg] off 
vertical to 22[deg] was widely supported. Several commenters viewed 
these changes as aligning the bench seat more with the ECE Regulation 
44 seat assembly bench, which has a seat back angle of 20 +/- 1[deg]. 
In response to commenters and in further consideration of the agency's 
efforts to harmonize its standards where possible, the agency amends 
the seat back angle by increasing it to 20 +/- 1[deg] to make it 
consistent with the test seat assembly of ECE Regulation 44. The agency 
believes that the difference between 22[deg] and 20 +/- 1[deg] is 
negligible and should have no significant effect on child restraint 
performance.
    ii. Belt Systems On The Standard Seat Assembly. The commenters 
generally agreed with the proposals for amending the seat belts on the 
test seat assembly. Almost all of the commenters supported increasing 
the spacing between the anchors of the lap belt from 222 millimeters 
(mm) to 392 mm in the center seating position and from 356 mm to 472 mm 
in the outboard seating positions. JPMA stated that it does not object 
to the proposal, but noted that the potential effect in side impact 
testing is unknown. Ms. Bidez suggested that the anchors should be set 
not at an averaged

[[Page 37631]]

spacing but at the maximum anchorage spacing ``now allowed'' for 
vehicle manufacturers in any seat position.
    This final rule adopts the proposals, except the spacing between 
the anchors of the lap belt in the center seating position will be 400 
mm, rather than 392 mm as proposed. The agency believes that the 8 mm 
difference between 400 and 392 mm is negligible, yet the 400 mm 
specification will make the spacing identical to that of the test seat 
assembly of ECE Regulation 44, so it is adopted. The lap belt anchorage 
spacing in the outboard seating position is revised to 472 mm, as 
proposed. (The ECE regulation specifies a spacing of 400 mm for both 
lap only tests and lap/shoulder tests. The agency cannot conclude that 
the difference between 472 mm and 400 mm is insignificant, so the 
agency is not adopting the ECE specification.) In response to Ms. 
Bidez, the Federal motor vehicle safety standards specify a minimum 
spacing for the anchorages, not a maximum. As to setting the anchorages 
at the maximum spacing that the agency has measured in its test 
program, the agency declines this suggestion. The agency does not have 
sufficient information to form the basis for a conclusion that a safety 
need exists to set the anchorages at the widest spacing observed on a 
vehicle seat. Further, setting the anchorages at the maximum spacing 
was not proposed in the NPRM or evaluated in the agency's test program 
at PAX River.
    A few commenters expressed some concerns about certain aspects of 
the test seat assembly's seat belts that were not addressed by the 
NPRM. GM, the Alliance, and ARCCA, Incorporated (ARCCA), stated that 
the seat belt lower anchors for both the center and outboard seating 
configurations do not represent typical anchorage locations found on 
new vehicles. As stated by the Alliance, ``The lap belt anchorages are 
too far back and too low and the lower anchors for the outboard seat 
are too high to represent a typical rear seat.'' GM and the Alliance 
also believed that the current two-piece lap and shoulder belt should 
be replaced with a three-point continuous loop shoulder/lap belt with a 
simulated retractor. Ford suggested that, to improve reproducibility of 
test results, the standard should specify a ``reasonably tight'' 
tolerance of 8% +/- 1% elongation at 10,000 N for the belt webbing used 
on the standard test bench.
    The agency did not pursue revising the fore-and-aft and vertical 
placement of the seat belt anchorages in response to the TREAD Act. 
This was due in part to the short deadlines of the TREAD Act. In 
addition, information from a 1994 test program indicated an absence of 
a need to change those anchorage locations. In 1994, the agency 
explored locating lap and shoulder belt anchorages on the standard seat 
assembly in a test program supporting rulemaking amending FMVSS No. 213 
to facilitate the production of belt-positioning booster seats. The 
agency found that the fore-aft and vertical placement of the lap belt 
had a negligible effect on the performance of the child restraints 
evaluated in the program. 59 FR 37167, 37171; July 21, 1994. 
Nonetheless, in that rulemaking the agency placed the inboard anchor to 
reflect the location of the average condition identified by the 
research. The agency believes that those fore-aft and vertical 
locations are still sufficiently representative of current vehicles so 
as to provide a true and thorough evaluation of a child restraint's 
performance in a crash.
    Given agency resources and rulemaking priorities, NHTSA does not 
anticipate exploring in the near future whether the fore-aft and 
vertical placement of the lap belt anchorages should be changed, or 
whether the current two-piece lap and shoulder belt should be replaced 
with a three-point continuous loop lap/shoulder belt with a simulated 
retractor. Our assessment of the safety need for such a rulemaking 
could change, if new information arises that indicates that these 
issues should be explored.
    In response to the issue raised by Ford, the elongation of the 
standard belt webbing used in FMVSS No. 213's compliance test was not 
discussed in the NPRM. It should be noted that specifying elongation of 
the webbing was addressed by NHTSA in the July 21, 1994 final rule on 
belt-positioning boosters (59 FR at 37171). Under current FMVSS No. 213 
test procedures, NHTSA tests child restraint systems using webbing that 
is typical of that installed in vehicles. NHTSA obtains webbing 
material from seat belt suppliers. These suppliers also furnish vehicle 
manufacturers with the webbing used in motor vehicles. This aspect of 
the compliance test increases the likelihood that the belts used to 
attach child restraints to the standard seat assembly are those that 
will actually be used by consumers to attach the restraints to their 
vehicle seats.
    The belt webbing is required by FMVSS No. 209 (S4.2(c)) to meet 
elongation requirements. Ford believed that the elongation allowed by 
that standard is too varied (``from zero to twenty percent for a lap 
belt, * * * up to 30 percent for the pelvic portion of a lap/shoulder 
belt, and * * * up to 40 percent for the upper torso portion of a lap/
shoulder belt. Such a large permitted variation in choice of belt 
webbing elongation could markedly affect FMVSS 213 dynamic test 
results.'') Ford did not provide data substantiating that differences 
in test results were obtained that were attributed to the use of 
webbing with different elongation characteristics. The agency also 
cannot conclude that testing with webbing with a ``tight tolerance'' of 
8 percent, as Ford suggested, is preferable over testing with webbing 
with a larger tolerance, e.g., closer to the 30 or 40 percent limit. 
Given agency resources and priorities, the agency can not conclude that 
a need exists to initiate rulemaking on this aspect of FMVSS No. 213 in 
the near future.
    iii. Fixed Seat Back. Commenters did not see eye-to-eye on the 
proposal to change the seat back to represent a fixed vehicle seat. 
Graco, TraumaLink, the Alliance, Safekids, Evenflo, JPMA and Xportation 
supported the proposal. JPMA stated that a fixed seat back replicates 
today's motor vehicle seat back and harmonizes with the test bench 
setups for ECE, Canadian and Australian regulations. Xportation said 
that it did not believe that motion of seat backs in vans is 
significant to the performance of child restraints. On the other hand, 
General Motors agreed with the proposal that a fixed seat back would be 
more representative of the rear seat of today's passenger cars, but 
expressed concern that a fixed back would not be representative of 
free-standing seats in vans and other multipurpose passenger vehicles. 
GM believed that it was unclear how fixing the seat back would affect 
child restraint system performance and suggested that NHTSA should 
study the issue. Advocates and Ms. Bidez expressed concern that 
changing to a rigid seat back may result in a less rigorous test 
condition, even though, the commenter believed, ``many children will be 
seated in seats with flexible seat backs.'' ARCCA believed that the 
configuration that results in the more severe test of a child restraint 
should be selected.
    In an effort to assure that the proposed fixed seatback 
configuration does not pose a less stringent test condition for dynamic 
tests of child restraints than the existing flexible seatback, NHTSA 
conducted a series of rigid versus flexible seatback tests at the 
agency's Vehicle Research and Test Center (VRTC) on September 23-27, 
2002. The proposed seatback and seat base angles were used.
    Six pairs of tests using rigid and flexible seatbacks were 
conducted using

[[Page 37632]]

the CRABI-12-month, and the Hybrid III 3- and 6-year-old dummies in 
rear- and forward-facing seat configurations, all with lap or lap and 
shoulder belt attachments (a top tether was not used). Charts providing 
plots of the normalized injury criteria measurements from these tests 
for HIC, chest acceleration and head and knee excursions are provided 
in the document titled, ``Comparison of Flexible and Rigid Seat Backs--
FMVSS No. 213 Test Assembly,'' which has been placed in the docket.
    The CRABI 12-month-old dummy was tested in a rear-facing infant-
only child restraint with both the rigid and the flexible seat backs. 
Charts A and B of the aforementioned document provide plots of the 
normalized injury criteria measurements from these tests for HIC and 
chest acceleration. There are no established head and knee excursion 
limits for rear-facing child restraints.
    The Hybrid III 3-year-old dummy was tested in three forward-facing 
child restraints--a 5-point harness, an overhead shield, and a shield-
type booster with the shield in place--using both the rigid and 
flexible seat backs as in the tests with the CRABI dummy. Charts C 
through K provide plots of the normalized injury criteria measurements 
from these tests for HIC, chest acceleration, head and knee excursion.
    Similar tests were conducted using the Hybrid III 6-year-old dummy 
in both a backless belt-positioning booster and in a high-back belt-
positioning booster seat. The plots of the normalized injury criteria 
measurements are provided in Charts L through Q of the document.
    In each of the tested configurations (e.g., 3-year-old dummy in an 
overhead shield convertible restraint), only one set of rigid versus 
flexible comparison tests was run. As such, the data used to evaluate 
the effects of the seat back are limited at best. The data were 
inconclusive as to whether a rigid seat back represents a less vigorous 
test. Review of the data indicates that, in some cases, the move to a 
rigid seat back resulted in a reduction in measured dummy response 
(lower HIC and chest g's for the 3-year-old dummy in overhead shield 
convertible). However, other cases show increases in dummy response 
when the rigid seat back is used (higher HIC for 3-year-old dummy in 5-
point harness convertible, shield booster; also for 6-year-old dummy in 
backless belt-positioning booster). Importantly, NHTSA notes that where 
differences in performance were noted for a particular injury criteria 
in a tested configuration, those differences were typically very small. 
Furthermore, in nearly each instance, results for both the rigid and 
the flexible configurations were within a 20 percent compliance margin 
indicating a level of performance that is well within the established 
limits.
    Based on the above data, NHTSA concludes that any differences seen 
between testing conducted with a rigid versus a flexible seat back 
would be minimal, and therefore, a move to a rigid seat back would not 
represent a less stringent test for child restraints. Further, the 
agency notes that there are more passenger cars (with rigid seat backs) 
than vans and multipurpose vehicles (with more flexible seat backs) in 
the existing vehicle fleet. As such, the move to a rigid seat back 
would more closely represent the existing vehicles on the road. The 
rigid seat back, on balance, will not be a less stringent requirement, 
and that it will allow child restraint performance optimization more 
representative of the vehicle fleet. In addition, a rigid seat back 
further harmonizes the standard's test seat assembly with ECE 
Regulation 44, which specifies a rigid seat back in testing child 
restraints to that standard. For the above reasons, NHTSA is adopting 
the rigid seat back as proposed in the NPRM.
    Figure 1A of FMVSS No. 213 is revised to reflect the above changes, 
as is the drawing package of the seat assembly that is incorporated by 
reference into the standard. (This final rule makes a technical 
amendment to 49 CFR 571.5 to provide information on obtaining copies of 
the drawing package).
    iv. Future Work. The agency tentatively decided in the NPRM that 
certain features of the bench seat need not be changed because they 
either reflected the design of production seats or are different but 
the difference was deemed not to have an effect on child restraint 
performance in dynamic testing.
    Seat Cushion Length: NHTSA found that the current FMVSS No. 213 
seat assembly has a seat pan length that is about 50 mm longer than the 
average seat pan length observed in today's vehicle fleet. The agency 
did not believe that the difference was significant. Commenters 
Consumers Union, Ms. Bidez, SafetyBeltSafe and ARCCA believed that the 
agency should consider shortening the length of the seat cushion to 
reflect a more demanding test condition. Ford commented that the 
current seat cushion is about the same length as a typical rear seat 
cushion, but suggested that the support for the seat cushion be 
extended to more realistically support the front edge of the cushion.
    NHTSA continues to believe that the length of the seat cushion of 
the standard seat assembly need not be changed, as it closely reflects 
production seats and because there is no information indicating that 
the difference in seat cushion length may affect child restraint 
performance on the seat. In addition, in view of the time constraints 
of the TREAD Act, NHTSA did not assess seat cushion support. However, 
the agency does not believe that seat support is critical. While some 
existing passenger cars will likely have a seat cushion that is 
supported more fully toward the leading edge of the cushion, vans and 
SUVs with bench-type seats that are removable or foldable, or 
individual seats such as ``Captains Chairs'' typically found in the 
second row of seating positions, will likely have much less support 
toward the leading edge of the seat cushion than in passenger cars. The 
agency does not anticipate undertaking efforts to evaluate which of 
these conditions would provide a more stringent test.
    Test Bench Floor: Graco and Ford indicated support for the addition 
of a floor onto the test bench for testing or rating \7\ child 
restraints. NHTSA does not believe that the standard seat assembly 
needs a floor because child restraints must meet the requirements of 
FMVSS No. 213 when attached to the seat assembly by use of the seat 
belts and LATCH system, without use of supplemental floor braces or 
other attachments. The commenters also suggest that an agency consumer 
information program rating the performance of child restraints should 
utilize all features with which the restraint is equipped, including 
those that are optional, i.e., that are not necessary for the restraint 
to meet Standard No. 213. The agency will consider the suggestion when 
developing its upcoming consumer information pilot programs relating to 
child restraint performance.
---------------------------------------------------------------------------

    \7\ Section 14(g) of the TREAD Act directed NHTSA to establish a 
child restraint safety ratings program. The agency has established 
an ease of use ratings program and will be conducting pilot programs 
on possible ratings programs geared toward rating child restraint 
performance in sled tests and vehicle performance in frontal vehicle 
crash tests. 67 FR 67491 (November 5, 2002)(Docket 02-10053).
---------------------------------------------------------------------------

    Seat Cushion Stiffness: The question of the stiffness of the seat 
cushion attracted most of the comments relating to features of the seat 
assembly that the NPRM did not propose to change. The NPRM stated that 
the agency was interested in increasing the stiffness of the cushion, 
but was uncertain what differences, if any, could be seen in

[[Page 37633]]

dynamic testing. Comments were requested on what the stiffness should 
be (67 FR at 21812).
    Several commenters believed that the stiffness of the seat cushion 
has a strong effect on child restraint performance. Consumers Union 
(CU) commented that it believed that cushion stiffness plays a major 
role in child restraint installation and suggested that further tests 
and analysis were needed. UMTRI expressed concern that the foam of the 
present test seat assembly is softer than many seats in the current 
fleet: ``Instead of representing a worst-case scenario, the response of 
the soft foam and its tendency to bottom-out on to the unrealistically 
stiff plywood backing can lead to misleading results that can reduce 
the level of child passenger safety.'' Ms. Bidez believed that cushion 
stiffness has a critical influence on child restraint performance 
relative to head excursion. These commenters did not provide supporting 
data.
    Some commenters were uncertain whether performance would be 
affected. JPMA stated that it conducted a small group of tests to 
evaluate the effect of foam in the tests, but the results ``yielded 
more questions than it answered.'' Without elaborating on its 
statement, JPMA provided data from a test program it conducted on foam 
that was 4 inches thick with a 25 percent compression/deflection 
resistance of 49.5 lb.\8\ The effect on the performance of test dummies 
in various types of child restraints was varied. JPMA stated that it 
did not believe that there is yet enough information to evaluate what 
the foam firmness and density should be, or how child restraint 
performance would be affected by changing the foam. In its own comment, 
Graco also expressed that it was unsure of how performance would be 
affected and suggested that testing and research be completed before 
changing the foam.
---------------------------------------------------------------------------

    \8\ The foam in the current seat assembly is thicker and softer 
than the foam tested by JPMA. The foam in the current seat assembly 
is 6 inches thick. Two pieces of foam (one 2 inches and one 4 
inches) may be used to achieve the required dimension. To be 
suitable for compliance testing, foam inserts must compress 25 
percent under the following load limits: (1) 45-55 pounds for the 2-
inch thick foam, and (2) 21-27 pounds for the 4-inch thick foam.
---------------------------------------------------------------------------

    Commenters had different views as to how the seat cushion foam 
should be changed. JPMA expressed cautious support for changing the 
foam to resemble more closely the foam thickness and compression of 
rear seats in real-world automobiles. UMTRI suggested that the agency 
characterize the overall seat stiffness of several modern vehicles and 
select a foam stiffness that matches a mean response. Ford stated that 
current rear seats are typically thinner and firmer than the test bench 
seat cushion. Ms. Bidez believed that the test cushion must reflect the 
softer seats of the majority of used vehicles on the road today. ARCCA 
believed that the seat cushion in Standard No. 213 may be too thick to 
match the vehicle seats, thereby allowing more deflection before 
becoming stiffer. The commenter suggested that the standard ``should 
err on the side of a softer cushion which will likely result in 
increased occupant excursion * * *.''
    After reviewing the comments and considering the agency's research 
needs and limited resources, NHTSA has decided not to endeavor at this 
time to change the stiffness of the standard seat assembly's seat 
cushion foam. As discussed in the NPRM, NHTSA is aware of data that 
indicate that the stiffness of the seat assembly cushion might not have 
a marked effect on child restraint performance. The agency conducted a 
study in 1988 comparing the stiffness characteristics of the seat 
assembly cushion with the characteristics of then current seats. 67 FR 
at 21812. Most vehicle seats were stiffer than the FMVSS No. 213 seat 
assembly. Sled tests were performed in the study to compare the dummy 
responses of the standard's seat cushion, a representative seat cushion 
that was softer, and a stiff cushion. The agency concluded that dummy 
response differences were not sufficiently large or consistent to 
warrant specifying a different cushion than that used in the current 
test seat assembly. Because possibly revising this parameter of the 
seat assembly would require further research, utilizing scarce agency 
resources, for disproportionate safety benefits, the agency will not 
pursue changing seat cushion stiffness for the time being.
    Harmonize With Transport Canada: Several commenters concurred with 
the NPRM that the proposed changes to the test seat assembly would 
advance harmonization with ECE Regulation 44 in that the seat cushion 
and seat back angles would be similar, as would the lateral spacing of 
the seat belt anchors and the rigidity of the seat back. However, the 
Alliance, General Motors and Evenflo noted that the test bench would 
differ from that used by Transport Canada in testing child restraints 
to the Canadian child restraint standard. These commenters urged NHTSA 
to work with Transport Canada to ensure that the test benches are 
harmonized.
    NHTSA regularly coordinates its vehicle safety plans and programs 
with Transport Canada and the agencies work closely on regulatory 
initiatives concerning child restraint safety. Harmonizing the 
countries' requirements to the extent consistent with the safety needs 
of each country is a goal shared by both entities. Specifically with 
respect to the TREAD Act, NHTSA has discussed each of the revisions 
with Transport Canada. Transport Canada is aware of the changes, and 
the agencies will continue efforts to harmonize regulations to the 
extent possible.

b. Crash Pulse

    The comments received on this aspect of the NPRM focused generally 
on the issues of the sled pulse shape (widening of the corridor) and 
severity.\9\
---------------------------------------------------------------------------

    \9\ Ford was concerned that the proposed pulse only specified 
sled movement during the first 90 ms, but limited dummy responses 
for 300 ms. Ford stated: ``Braking of a Hyge sled can have a 
substantial effect on dummy kinematics and readings during rebound. 
Hyge sled tests are generally considered to be unrealistic during 
the rebound phase because of sled braking. If the agency believes 
that it is essential to limit dummy measurements during rebound, and 
the agency plans to use a Hyge-type sled for audit testing, sled 
accelerations between 90 and 300 ms should be limited to specify an 
objective test pulse.'' The agency does not agree that sled braking 
has caused objectivity problems in the past. The FMVSS No. 208 sled 
test (see Figure 6 of that standard) specifies a sled corridor only 
to 130 ms, but at least 300 ms of data is collected in measuring 
injury criteria. There have not been any problems with the effect of 
the braking of Hyge sleds on dummy kinematics and readings during 
rebound. Accordingly, the agency is not specifying a pulse corridor 
between 90 and 300 ms.
---------------------------------------------------------------------------

1. On Widening the Corridor
    As for widening the corridor of the sled pulse from 80 milliseconds 
(ms) to approximately 90 ms in duration, all but few of the commenters 
responding to this issue supported the change. Many agreed with the 
agency that the change would allow more laboratories to run the 
compliance test ``without decreasing the effectiveness of the testing'' 
(quoting UMTRI). SafetyBeltSafe (SBS) also agreed with NHTSA's 
assessment, explained in the preamble to the NPRM, that the pulse would 
enable tests to be conducted closer to 30 mph.
    The JPMA and Graco did not support revising the corridor. JPMA 
stated that widening the corridor necessarily makes the standard more 
stringent, because child restraint manufacturers will have to design 
products that can comply at the new extremes of the compliance 
corridor. The commenter stated that difficulties experienced by test 
labs in fitting their pulses within the existing corridor ``should be 
addressed by insisting that the test labs figure out how to meet the 
existing test corridor.''

[[Page 37634]]

JPMA and Graco believed that a wider test corridor will necessarily 
lead to more lab-to-lab variability during certification and compliance 
testing, which, the commenters stated, increases the compliance burden 
on manufacturers. JPMA stated that the agency did not provide data on 
the effect of the different crash pulse with the new bench seat, and 
believed that the agency must assess the effect of a wider sled pulse 
corridor on child restraint compliance.
    The agency responds by concurring that the revision to the pulse 
could affect the manufacture of child restraints. Widening the test 
corridor from 80 ms to approximately 90 ms in duration does enable 
NHTSA to test child restraints closer to 30 mph than the present. To 
the extent that the 30 mph tests are more stringent than tests 
conducted in the past at slightly lower speeds, that result is a 
desired outcome of the amendment. Widening the corridor improves the 
effectiveness of the test. Child restraint manufacturers will have to 
certify that their child restraints meet the requirements of FMVSS No. 
213 when tested using the test pulse, possibly at a higher velocity. 
They may have to conduct some testing to ensure that the restraints can 
be certified to the requirements when tested in the more effective 
manner using this pulse. The agency acknowledged in the NPRM the likely 
need for manufacturers to retest their restraints because of the new 
seat assembly and, by implication, due to the changes to the crash 
pulse (67 FR at 21829). However, the agency believed then and continues 
to do so now that it is unlikely that child restraints must be 
redesigned because of the change in the assembly and pulse.\10\ 
Restraints are generally manufactured with enough of a compliance 
margin that will allow them to meet the requirements of the standard 
when tested at a slightly higher velocity.
---------------------------------------------------------------------------

    \10\ Note that the agency is not specifying a ``new'' crash 
pulse. Rather, the final rule puts a corridor specification around 
the existing pulse which allows the agency to conduct compliance 
tests at velocities closer to 30 mph.
---------------------------------------------------------------------------

    To illustrate, NHTSA examined some of the work that was performed 
in support of the development of the child restraint ratings program 
required under Section 14(g) of the TREAD Act. As part of this effort, 
the agency examined the margin by which existing child restraints meet 
the injury limits currently specified in FMVSS No. 213. In model year 
2000, the agency tested 50 upright, forward-facing child restraints in 
accordance under the agency's FMVSS No. 213 compliance test program. 
Twenty-four (24) seats were tested without a top tether, and 26 seats 
were tested with a top tether. We secured all seats with only a lap 
belt (no lower anchorages or shoulder belts). Currently, to pass the 
FMVSS No. 213 compliance test, a child restraint must achieve dummy 
injury numbers of a HIC less than 1,000 and a resultant chest 
acceleration of less than 60 G's. As shown below in Figure 1, 
regardless of whether we equipped the child restraints with a top 
tether, all child restraints achieved dummy injury readings below the 
maximum allowable values. Figures 2 and 3 illustrate the margin of 
compliance for HIC and chest acceleration, respectively. The margin of 
compliance is one minus the measured injury reading divided by the 
injury assessment reference value (IARV) times 100. Higher percentages 
are better, having less probability of injury. Regarding the HIC, all 
model year 2000 child restraints tested easily fall within the limits 
specified by the FMVSS No. 213 compliance tests. Most had a compliance 
margin of more than 50%. Although the margin is not as large for chest 
acceleration, all tested child restraints passed this compliance 
requirement as well.
BILLING CODE 4910-59-P

[[Page 37635]]

[GRAPHIC][TIFF OMITTED]TR18JN03.006


[[Page 37636]]


[GRAPHIC][TIFF OMITTED]TR18JN03.007


[[Page 37637]]


[GRAPHIC][TIFF OMITTED]TR18JN03.008

BILLING CODE 4910-59-C
    FMVSS. No. 213 also has a requirement for head and knee excursions. 
Head excursion is limited to 720 mm (28 in) when a top tether is used, 
and 813 mm (32 in) without use of a top tether. Knee excursion is 
limited to 915 mm (36 in). Figures 4 and 5 below illustrate the margin 
of compliance for head excursion and knee excursion, respectively. Head 
and knee excursion limits are compliance limits imposed to reduce the 
chances of a child striking the vehicle interior or submarining 
(sliding under the belt feet first) in an automotive crash. Head and 
knee excursions are much closer to the compliance limits than HIC and 
chest acceleration. This may reflect attention to occupant protection, 
since increases in distance traveled by the occupant reduces the forces 
experienced by the occupant.
BILLING CODE 4910-59-P

[[Page 37638]]

[GRAPHIC][TIFF OMITTED]TR18JN03.009


[[Page 37639]]


[GRAPHIC][TIFF OMITTED]TR18JN03.010

BILLING CODE 4910-59-C
    During the development of the child restraint ratings program, the 
agency also conducted dynamic testing of a number of child restraints 
both at 30 and 35 mph to examine what differences, if any, resulted 
from the increase in the velocity at which the test was conducted. To 
attain the higher speed, a sled pulse with a similar shape and duration 
length as that of the FMVSS No. 213 pulse was used, except that the 
change-of-velocity was elevated from 30 mph (48km/h) to 35 mph (56km/
h). All of the child restraints tested produced dummy injury 
measurements well below the FMVSS No. 208 criteria of 570 HIC and 55g 
chest acceleration (Hybrid III 3-year-old dummies were used in the 
tests). Although the injury assessment values were slightly greater in 
the 35 MPH (56 km/h) sled tests than in 30 mph (48 km/h) sled test, 
eight of the nine child seats tested rated within the 5 star range, and 
one fell just marginally below in the 4 star range. This data, in 
conjunction with the information provided above regarding the 
compliance margin achieved by existing child restraints, demonstrates 
that a nominal increase in the test velocity resulting from the crash 
pulse corridor established as part of this final rule will not 
necessitate a redesign of existing child restraint designs to meet the 
injury criteria limits established in the standard.
    The agency also does not believe that unusual or unacceptable 
variability will be introduced into the test results simply because 
more test labs will be involved in conducting child restraint tests. 
Any lab-to-lab variability resulting from a properly conducted test 
will be insignificant, in part because each laboratory must ensure that 
the pulse it uses in the FMVSS No. 213 sled test falls within the 
corridor specified in the standard. In addition, it is the 
responsibility of manufacturers to design and manufacture child 
restraints to meet the requirements of the standard, taking into 
account whatever variability occurs from seat-to-seat manufacturing 
differences and from lab-to-lab testing differences. It should also be 
noted that child restraint manufacturers are responsible for ensuring 
that their restraints meet the requirements of the standard when tested 
by NHTSA in its compliance test. Manufacturers testing their products 
to the most demanding requirements under the most demanding test 
conditions increase the likelihood that their products will meet the 
requirements when tested by NHTSA under the same or less severe 
conditions. In the same manner, prudent testing by the manufacturer 
accounts for routine lab-to-lab variability that may occur when testing 
child restraints. Manufacturers must design and produce products that 
will pass the compliance test regardless of the laboratory conducting 
the test.
2. Increase Pulse Severity
    ARCCA opposed the NPRM based on concerns that the proposed changes 
to the crash pulse would ``lower, rather than raise, the bar for child 
restraints.'' The commenter believed that the Standard No. 213 pulse is 
actually less severe than all of the 30 mph barrier test pulses from 
actual vehicles, and that the standard's pulse severity should be 
increased. The commenter suggested that the standard specify that the 
dynamic test will be conducted at velocities of not less than 30 mph. 
``This will ensure that manufacturers do not take advantage of the 
wider corridor to conduct testing that is less severe than what is 
currently required by FMVSS 213.'' ARCCA also stated that the standard 
``should contain a minimum acceptable peak acceleration level that is 
more than the 19 G's or [sic] the

[[Page 37640]]

proposed corridor in the NPRM.'' ARCCA stated:

This minimum acceleration level should be high enough to ensure that 
a child restraint will offer acceptable performance and be capable 
of remaining structurally intact. Testing performed by one auto 
manufacturer in a minivan demonstrated that various child restraints 
structurally failed in 30 mile per hour sled testing using the 
vehicle's barrier crash pulse. By setting a high minimum peak 
acceleration, confidence can be gained in the ability of a child 
seat to remain structurally intact and protect a child no matter in 
what vehicle it is installed.

    ARCCA suggested that the agency specify in Standard No. 213 that 
the test pulse must fall within a specific corridor and must have a 
velocity of at least 30 mph and a peak acceleration of at least some 
predetermined value. ARCCA believed that that acceleration value should 
be based on the values obtained from barrier crash tests and be greater 
than the majority of all FMVSS No. 208 tests reported. ARCCA was also 
concerned about how the values presented in Table 4 of the NPRM were 
calculated, especially the peak g values. The commenter believed that 
the values in the NPRM were erroneously based on ``average pulses'' 
i.e. point-by-point averaging of the pulse data to form a single curve 
for a class of vehicles. ARCCA stated that the problem with this method 
is that when pulses with peaks at different times are combined, the 
resulting peak is less than either of the pulses averaged. ``This is 
due to the fact that the crash pulses are out of phase. This is similar 
to the principle used in noise cancellation devices, when two waves are 
superimposed the magnitude of the resulting pulse is less.''
    The agency does not agree with ARCCA that the standard's pulse is 
deficient and should be increased. The pulse is representative of a 
severe crash and subjects child restraints to ``worst case'' testing in 
a sufficient manner. The severity of a crash pulse is determined 
through a combination of three factors: the acceleration onset rate, 
the peak acceleration, and the time duration of the pulse. The data 
presented in the PAX report are based on FMVSS No. 208 rigid barrier 
testing at 30 mph impact speed (approximately 32 mph total change in 
velocity, [Delta]V) and New Car Assessment Program (NCAP) rigid barrier 
testing at 35 mph (approximately 37 mph [Delta]V).
    The FMVSS No. 213 pulse was very similar to the pulses generated by 
sport utility vehicles (SUVs), trucks and small school buses in an 
FMVSS No. 208 (32 mph [Delta]V) crash test. NHTSA believes that the 
pulse should be severe enough to be adequately representative of these 
vehicles since child restraints are regularly and increasingly used in 
these types of vehicles. That is, the stringency of the pulse is 
justified to better ensure that each child restraint will not have 
structural degradation in a crash and will limit forces to the child's 
head, neck and torso to tolerable levels, no matter the vehicle the 
child is in.
    ARCCA was correct that the agency had averaged the pulses for the 
three classes of vehicles (SUVs, trucks and a small school bus) to 
develop a composite pulse for each vehicle class, and that the 
composite pulses had peak acceleration levels that are typically lower 
than the highest peak accelerations measured in the individual tests. 
However, the averaged pulses allowed the agency to examine general 
trends with respect to the crash parameters that determine the 
performance of vehicles in a crash. As such, they are representative of 
the pulses of vehicles in which child restraints are likely to be used 
and provide a reasonable foundation upon which the standard's pulse can 
be based. Further, the agency is unaware of the testing to which ARCCA 
referred that allegedly demonstrated ``that various child restraints 
structurally failed in 30 mile per hour sled testing using the 
vehicle's barrier crash pulse.'' To the contrary, child restraints have 
proven very effective in real world crashes and have performed well in 
the agency's studies of child restraint performance in vehicles tested 
in NCAP 35-mph frontal crashes.
    ARCCA suggested that the standard specify that the dynamic test 
will be conducted at velocities of not less than 30 mph. This 
specification is unnecessary, since the standard currently requires the 
dynamic tests to be frontal barrier impact simulations ``at a velocity 
change of 48 km/h [30 mph] with the acceleration of the test platform 
entirely within the curve shown in Figure 2 * * *.'' Thus, the agency 
already conducts the dynamic test at velocities as close as possible to 
30 mph without exceeding 30 mph or causing the pulse to fall outside of 
the curve of Figure 2 of the standard.
    ARCCA believed that the velocity of the sled test should be 
increased from 30 mph to 33 mph to replicate the change in velocity 
typically seen in a 208 barrier test. ``For the 213 pulse to be near 
the 30 mph barrier test the velocity, acceleration and duration would 
all have to be increased.'' The commenter also believed that, since 
``well-restrained adult occupants are capable of surviving crashes 
comparable to a 35 mph barrier crash where the change in velocity is 
closer to 40 mph,'' tests of child restraints should be performed at 
the levels specified by the agency in testing vehicles in the New Car 
Assessment Program.
    In contrast, all other commenters except ARCCA commenting on this 
issue did not want to increase the severity of the crash pulse. 
SafetyBeltSafe (SBS) believed that the velocity change should not be 
raised to 33 mph because ``the current test is already reflective of 
the top few percent of crashes.'' SBS stated that increasing the 
velocity ``will not significantly improve child restraint performance 
in the real world but will surely make the products more expensive.'' 
Graco stated that if the pulse were increased to 33 mph, it would 
expect a large number of child restraints needing to be redesigned with 
``minimal benefit to child passenger safety.'' UMTRI stated that the 
change in velocity for the test should remain at 30 mph, stating that 
it conducted a recent analysis of National Automotive Sampling System 
(NASS) data from 1995-2000 which showed that a 30 mph change in 
velocity is more severe than approximately 98 percent of the frontal 
impact crashes nationwide. UMTRI further noted that since the NASS 
database only includes tow-away crashes, ``this is a conservative 
estimate of the percentage of frontal impacts that are less severe than 
30 mph.'' UMTRI was concerned that increasing the velocity of the test 
is not likely to increase safety, but will increase consumer cost of 
child restraints and may lead to child restraint designs that could 
make the restraints less effective or more easily misused at lower 
severity crashes, ``which occur much more frequently.'' The Insurance 
Institute for Highway Safety (IIHS) stated that its review of NASS 
cases showed that child restraints designed to pass the current 30 mph 
sled test are providing very good protection to children in frontal 
crashes. IIHS also stated, ``There was no indication, based on an 
analysis of injuries, crash description, and photos in these 10 frontal 
crashes that designing child restraints to withstand higher crash 
forces could have prevented or mitigated any of the serious or fatal 
injuries.''
    NHTSA concurs with these comments that the standard's crash pulse 
adequately meets a safety need. Increasing the severity could 
necessitate the redesign of many child restraints and could increase 
costs of the restraints to manufacturers, without a proportionate 
safety benefit. Thus, the agency concludes that the pulse should not be 
made more severe at this time.

[[Page 37641]]

3. Decrease Pulse Severity
    While there was almost unanimous agreement among commenters that 
the crash pulse should not be increased, commenters expressed opposing 
opinions on whether the severity of the test pulse should be decreased. 
The crash pulse is more severe than most other pulses, but is similar 
to crash pulses of large sport utility vehicles and light trucks 
(passenger vehicles that are becoming more and more popular for use as 
family vehicles) and very similar to the crash pulse of small school 
buses. The agency determined in the NPRM that the crash pulse should 
maintain its level of stringency so as to replicate vehicle crashes 
involving vehicles that had relatively severe crash pulses. Some 
commenters disagreed, believing that the crash pulse should be reduced 
in severity because the most frequent crashes involving children in 
child restraints are those with lower crash pulse severities than the 
test pulse, while others agreed that a relatively severe, ``worst 
case'' scenario should be replicated.
    In support of reducing the severity of the crash pulse, the 
Alliance of Automobile Manufacturers (Alliance) stated that the current 
sled pulse represents--

an extremely rare ``worst case'' [(e.g., a stiff vehicle hitting a 
full-width non-deformable wall at high speed)]. As a result the 
addition of the new dummies/injury criteria coupled with this 
unrepresentative test pulse may create significantly unintended 
consequences such as reduced availability and increased costs of 
compliant restraints as well as the addition of features that may 
make them more cumbersome and less user friendly. All of which will 
reduce their use in the real world.

    The Alliance stated that an attachment it submitted with its 
comment contains an analysis comparing the severity (acceleration 
pulses) of full frontal barrier crashes with vehicle-to-vehicle crash 
tests. ``In this analysis a 30 mph full frontal barrier test is found 
equivalent to a 41 mph vehicle-to-vehicle crash. A reduced speed of 22 
mph for full frontal rigid barrier test is found to represent vehicle-
to-vehicle crashes with 50%-100% overlap, with each vehicle traveling 
at 30 mph.''
    Along the same lines, General Motors (GM) believed that the crash 
pulse should represent the most frequent collision event. The commenter 
urged research to define the real world collision speeds and 
deceleration pulses at which the majority of the harm to children 
occurs. GM believed that increasing the pulse duration and widening the 
corridor increases the pulse severity somewhat, and coupling this 
increase with the use of the new test dummies and injury criteria 
``could make compliance more difficult.'' GM suggested that NHTSA 
consider using the FMVSS No. 208 generic sled pulse if the final rule 
adopts the Hybrid III test dummies and injury measures proposed in the 
NPRM.
    The Children's Hospital of Philadelphia (TraumaLink) supported 
altering the pulse to be more representative of the passenger car 
environment to ``make it more relevant to a larger proportion of the 
real-world crash-involved population.'' The commenter stated that out 
of the 59,968 children studied in TraumaLink's Partners for Child 
Passenger Safety study, only 24.1 percent of children were riding in 
SUV's and light trucks.
    In contrast, in support of the agency's decision not to reduce the 
severity of the crash pulse, Advocates for Highway and Auto Safety 
(Advocates) believed that although cars remain more numerous in the 
vehicle fleet, use of an LTV crash pulse is representative of real-
world crash experience given that increasing numbers of LTVs have 
entered the fleet and are frequently used as passenger and family 
vehicles. The commenter also discussed why it believed the crash pulse 
should replicate the ``worst case'' scenario over the ``most frequent'' 
or ``average'' crash:

    Although Advocates has urged the agency to update its test 
procedures in certain respects to ensure that they are 
representative of the modern vehicle fleet, this does not mean that 
critical test procedures should mirror the attributes or test the 
performance of only the ``average'' vehicle. While test procedures 
should be representative of the vehicle fleet in many respects, not 
all tests or test procedures should be based on the most common or 
average vehicle in the fleet. To ensure safety protection for all 
vehicle occupants, critical aspects of test procedures should 
replicate more stringent conditions than would be experienced in the 
average vehicle. This is especially true when only one test and a 
single set of test conditions are used as the basis for compliance. 
Thus, although there are still more cars than LTVs on U.S. highways, 
and even though more children are injured while riding in cars than 
are injured while riding in LTVs, the FMVSS 213 sled test should 
replicate the faster acceleration onset rate and higher peak 
acceleration exerted in an SUV crash pulse. Of the two, the LTV 
crash pulse presents the more stringent test condition. Using the 
LTV-like crash pulse ensures that children exposed to such a severe 
force, as well as children exposed to less severe conditions in 
cars, will be afforded protection. The reverse, however, is not 
true. If FMVSS 213 adopted a car-like sled test crash pulse, 
children in cars may be protected but that same degree of safety 
would not necessarily be provided to children in LTVs with 
``stiffer'' frames that transfer more of the crash generated energy 
to the occupants. As a result, Advocates concurs in the agency's 
judgment that the existing FMVSS 213 crash pulse be retained.

    After reviewing all the comments on this issue, NHTSA has decided 
to retain the current severity of the pulse and not reduce it. The 
agency concurs with Advocates that to ensure safety protection for as 
many child occupants as possible, ``critical aspects of test procedures 
should replicate more stringent conditions than would be experienced in 
the average vehicle,'' and that, given that child restraints are used 
with a wide range of vehicle types and are involved in crashes of 
varying degrees of severity, such a critical aspect is the sled pulse. 
Accordingly, the agency declines to replicate the crash conditions of 
the most frequent collision event.
    GM suggested that NHTSA consider using the FMVSS No. 208 generic 
sled pulse if this final rule adopts the Hybrid III test dummies and 
injury measures proposed in the NPRM. As discussed later in this 
preamble, this final rule adopts the Hybrid III test dummies but does 
not adopt the majority of the injury measures proposed in the NPRM. 
Nonetheless, the agency makes the following observations about the 
suggestion to use the FMVSS No. 208 generic sled pulse. The generic 
sled pulse is less severe than the FMVSS No. 213 pulse. As shown in the 
following overlay of the existing FMVSS No. 213 pulse with the FMVSS 
No. 208 generic sled pulse, the former has a greater onset rate, higher 
peak acceleration and shorter time duration. Further, the FMVSS No. 208 
sled pulse, with a peak acceleration of about 17 g's, is less stringent 
than most 30 mph passenger vehicle crashes. Because the FMVSS No. 208 
sled pulse is less severe than the FMVSS No. 213 pulse, this final rule 
declines the suggestion to adopt it.
BILLING CODE 4910-59-P

[[Page 37642]]

[GRAPHIC][TIFF OMITTED]TR18JN03.011

BILLING CODE 4910-59-C

[[Page 37643]]

c. New Dummies

1. Post-NPRM Test Program
    As part of the test program conducted for NHTSA at the Patuxent 
River (PAX) test center, PAX conducted a series of dynamic sled tests 
to evaluate identical child restraints on the revised test seat 
assembly using both the Hybrid II and the Hybrid III 3- and 6-year-old 
dummies. All of these tests were conducted with the restraints attached 
to the test seat assembly with the lap belt only, as would be done in a 
compliance test. Similar comparison tests were conducted with the 
Hybrid II 9-month-old and the CRABI 12-month-old dummy, but as the 9-
month-old dummy is uninstrumented, little comparative information was 
gleaned from these tests. Accordingly, the data from the latter tests 
are not provided.
    i. Tests With The 3-Year-Old Dummies. The following Table 6 
illustrates the injury criteria measurements for the test series using 
the Hybrid II and Hybrid III 3-year-old dummies:

                                              Table 6.--Tests With the Hybrid II and Hybrid III 3-Year-Olds
--------------------------------------------------------------------------------------------------------------------------------------------------------
                                                                             HICunlimited     Chest acceleration    Head excursion      Knee excursion
              Child restraint                           Dummy            -------------------------------------------------------------------------------
                                                                            Value    Change     Value    Change     Value    Change     Value    Change
--------------------------------------------------------------------------------------------------------------------------------------------------------
Cosco Touriva.............................  Hybrid II...................     702.8  ........      40.4  ........      19.6  ........      26.4  ........
                                            Hybrid III..................     446.8      -256      37.6      -2.8      15.5      -4.1      26.4         0
Century Accel.............................  Hybrid II...................     626.5  ........      26.8  ........      19.5  ........      26.8  ........
                                            Hybrid III..................     355.3    -271.3      36.1      +9.3      19.9      +0.4      25.2      -1.6
Century Breverra..........................  Hybrid II...................     669.7  ........      29.2  ........      22.5  ........      27.4  ........
                                            Hybrid III..................     536.8    -132.9      50.1     +20.9      21.3      -1.2      29.1      +1.7
Cosco HB Booster..........................  Hybrid II...................     446.4  ........      41.6  ........      22.5  ........        26  ........
                                            Hybrid III..................     704.9    +258.5      41.6         0      13.4      -9.1      22.4      -3.6
--------------------------------------------------------------------------------------------------------------------------------------------------------

    The Cosco Touriva and the Century Accel are both forward-facing 
convertible child restraints, and the Century Breverra and the Cosco 
High Back Booster are forward-facing hybrid boosters. All were tested 
with the dummy in the restraint's internal harness system.
    The results from this series of testing appear to be mixed. Three 
of four tests showed a marked decrease in measured HIC values when 
testing with the Hybrid III dummy as compared to the Hybrid II dummy, 
while the fourth test in the series resulted in a significant increase 
(446.4 to 704.9) in HIC values. Similar results are seen when looking 
at chest acceleration and head and knee excursions. The varied results 
can be attributable in part to the very limited sample size of child 
restraints tested. No repeatability tests were performed. All injury 
numbers were well within the current limits prescribed in FMVSS No. 
213.
    ii. Tests With The 6-Year-Old Dummies. A similar series of tests 
was conducted with the Hybrid II and Hybrid III 6-year-old dummies in 
both backless and high back belt-positioning booster seats on the 
revised test seat assembly. As was the case in tests with the 3-year-
old dummies, the test results for the 6-year-old dummies show 
considerable fluctuation and no clear trends.
    The following Table 7 outlines the results of these tests:

                                              Table 7.--Tests With the Hybrid II and Hybrid III 6-Year-Olds
--------------------------------------------------------------------------------------------------------------------------------------------------------
                                                                                  HIC         Chest acceleration    Head excursion      Knee excursion
              Child restraint                           Dummy            -------------------------------------------------------------------------------
                                                                            Value    Change     Value    Change     Value    Change     Value    Change
--------------------------------------------------------------------------------------------------------------------------------------------------------
Cosco Gr. Explorer........................  Hybrid II...................     267.1  ........      49.2  ........      14.3  ........        24  ........
Evenflo Right Fit.........................  Hybrid III..................     357.6     +90.5      37.8     -11.4      11.7      -2.6      19.6      -4.4
Cosco Gr. Explorer........................  Hybrid II...................     328.2  ........      38.6  ........        18  ........      25.7  ........
Evenflo Right Fit.........................  Hybrid III..................     276.2       -52        36      -2.6      19.1      +1.1        21      -4.7
Century Breverra..........................  Hybrid II...................     209.4  ........      35.1  ........      19.7  ........      19.7  ........
                                            Hybrid III..................     415.7    +206.3      41.4      +6.3        20      +0.3      11.7      -8.0
Cosco HB Booster..........................  Hybrid II...................     380.7  ........      42.4  ........      17.6  ........      27.6  ........
                                            Hybrid III..................     756.1    +375.4      38.3      -4.1      18.4      +0.8        24      -3.2
--------------------------------------------------------------------------------------------------------------------------------------------------------

    The original test matrix called for testing each restraint with 
both the Hybrid II and the Hybrid III 6-year-olds to facilitate a 
direct comparison between the two dummies, as was done for the other 
dummies. However, during the conduct of the sled tests at PAX, the 
Cosco Grand Explorer was instead inadvertently tested twice with the 
Hybrid II 6-year-old, and the Evenflo Right Fit was tested twice with 
the Hybrid III 6-year-old.
    NHTSA acknowledges that this makes a direct comparison between the 
two dummies in the same restraint impossible. However, unlike rear-
facing infant seats and forward-facing toddler seats, there is very 
little difference in design characteristics between the two backless 
booster seats in question that would influence the dynamic response of 
the dummies in a sled test. As such, NHTSA has included the data for 
information.
    Further, it is noted that VRTC conducted a study comparing the 
performance between the Hybrid II and the Hybrid III child dummy 
families in support of the NPRM for this final rule. (See Docket NHTSA-
2002-11707-1; report dated April 12, 2002.) The report concluded in 
part that sled test results generally show fairly consistent dummy 
performance with the Hybrid II and Hybrid III child dummies.
2. Commenters Generally Supportive
    Commenters generally supported using the CRABI 12-month-old and the 
Hybrid III 3-year-old dummies in Standard No. 213 compliance tests, in 
place of the TNO 9-month and the

[[Page 37644]]

Hybrid II 3-year-old dummies now used by the agency. There was support 
for the use of the Hybrid III 6-year-old dummy in compliance tests, 
with the exception of a few commenters (discussed below). There was 
general concern about the need for and capabilities of the weighted 
Hybrid III 6-year-old dummy.
    i. Hybrid III 6-Year-Old Dummy. Several commenters expressed 
concern about the biofidelity of the unweighted Hybrid III (HIII) 6-
year-old dummy's neck and hips and the suitability of the dummy for use 
in testing child restraints. TraumaLink stated that, based on a sled 
test program it conducted at a test lab, they had ``significant 
concerns'' regarding the performance of the dummy. ``The tests revealed 
extremely large neck elongation unlikely to be seen in real children in 
real crashes and resulted in high calculated injury values. These 
results suggest a pattern of injuries that we do not see in our real 
world experience.''
    SafetyBeltSafe referred to the tests performed by TraumaLink to 
conclude that ``We do not now believe that the HIII 6-year-old dummy is 
an appropriate test device to simulate a restrained child'' because of 
``unrealistic stretching and bending of this dummy's neck while tightly 
restrained by a lap-shoulder belt in a booster. The result was that the 
dummy's face directly contacted the chest, generating an unrealistic 
and unacceptably high HIC.'' SafetyBeltSafe also stated that test data 
from NHTSA's Vehicle Research and Test Center indicated that tests with 
the dummy generated ``[head] excursion increases of from 2.1 to 4.5 
inches in a booster with lap-shoulder belt. The likely reason for this 
is that the neck is not a true Hybrid III type neck, as it lacks the 
metal disks needed to limit its bending.'' The commenter was also 
concerned about the dummy's ``permanently flexed hips, which, unlike 
the new 10-year-old design, do not allow a slouched position and may 
inhibit submarining in non-optimal booster designs.''
    Ford Motor Company likewise stated that the Hybrid III dummies are 
much more likely to experience head-to-knee contacts than Hybrid II 
dummies, because of the more flexible ribs and neck of the HIII 
dummies. Further, Ford said that a 1984 study (Culver et al.) showed 
that adult HIII dummy HIC readings were about twice those recorded in 
dummy head to cadaver knee impacts. Further, Ford stated that because 
the HIII 6-year-old dummy does not have the metal plates that segment 
and limit bending of the necks of the HIII adult dummies, the HIII 6-
year-old dummy may be more likely to experience head-to-leg contacts 
than ``the three-year-old.'' Ford asked in its comment how the agency 
would treat head acceleration spikes that could be caused by head-to-
knee contacts. The commenter also suggested that load cells be used on 
the ASIS of the pelvis of the 6-year-old dummy to evaluate the tendency 
to submarine under the lap belt during testing of booster seats, 
because, Ford stated without elaborating, the current limit on knee 
excursion is not an effective way to limit submarining in tests of 
belt-positioning boosters.
    NHTSA disagrees with the commenters that the HIII 6-year-old dummy 
should not be used in FMVSS No. 213 testing. The neck of the HIII 6-
year-old is currently performing within the specifications established 
by the Hybrid III Dummy Family Task Force of the Society of Automotive 
Engineers (SAE). The agency is not aware of specific test information 
and/or data substantiating the claims of the commenters that the dummy 
is an unsuitable test device for FMVSS No. 213 testing.
    When the dummy was incorporated into the regulation on 
anthropomorphic test devices, 49 CFR part 572, the agency made the 
following determinations (65 FR 2059) about the dummy:

Based on NHTSA's use of the H-III6C 6-year-old dummy in calibration 
tests and in frontal impact tests involving restraints such as air 
bags and belts, we have concluded that this dummy is suitable for 
both research and compliance safety assessments. The dummy is not 
only considerably more biofidelic than its predecessor, the part 572 
subpart I 6-year-old dummy, but it also has considerably more 
extensive instrumentation to measure impact responses such as 
forces, accelerations, moments, and deflections in conducting tests 
to evaluate vehicle occupant protection systems.

    The agency continues to believe that the performance of child 
restraint systems will be more thoroughly and precisely assessed by use 
of the HIII dummy because of the dummy's enhanced biofidelity and 
extensive instrumentation. With regard to concerns about the dummy's 
neck, it should be noted that the Hybrid II dummy currently in use also 
does not have the metal disks. Since the Hybrid III is more 
biomechanically based, we continue to believe that it provides a more 
humanlike response than the Hybrid II version of the dummy.
    Sled tests have shown the HIII 6-year-old to be a suitable 
replacement for the existing HII 6-year-old in FMVSS No. 213 compliance 
tests. None of the sled testing conducted with the HIII 6-year-old 
dummy at VRTC or PAX in support of the TREAD Act has indicated that 
head-to-chest or head-to-knee impacts is an issue. Such impacts are not 
typical.\11\ NHTSA believes that if head-to-knee contact occurs, there 
are likely design concerns with respect to the particular child 
restraint that should be addressed to eliminate such contact. We also 
believe it would be very difficult, if not impossible, to establish an 
objective means to determine if, and if so to what extent, head-to-knee 
contact influenced HIC measurement in FMVSS No. 213 compliance testing. 
Consequently, head acceleration spikes caused by head-to-knee contacts 
will be included in the HIC computation. Further, the agency continues 
to believe that the HIII dummy is needed to better assess the injury 
mechanisms to children.
---------------------------------------------------------------------------

    \11\ The agency is aware of only one instance in which there was 
significant head-to-knee contact in an FMVSS No. 213 test 
environment using a Hybrid III dummy. In this case, a 6-year-old 
dummy was tested in a backless belt-positioning booster. In the 
test, the shoulder portion of the belt system slipped off the 
dummy's shoulder. It is unclear what caused this to happen.
---------------------------------------------------------------------------

    The agency is not entirely convinced that neck elongation is not 
occurring to children in real crashes. We believe it possible that neck 
injury may sometimes not be diagnosed even though it occurs. Since a 
child's neck is not fully developed, detection of injuries is more 
difficult and injuries could manifest in later years. Also, for fatal 
injuries, there is often a reluctance to conduct autopsies in deference 
to family sensitivity. Consequently, the cause of death may be listed 
as massive head injury, while injury to the neck may have also 
occurred.
    The agency is continuing to conduct research to establish better 
neck injury response and injury criteria for children. Research may 
show the presence of neck injury and a possible need for a neck injury 
criterion in FMVSS No. 213. If that occurs, a test dummy incorporated 
into the standard that offers improved biofidelity and neck 
instrumentation would prove useful. Because we believe that the current 
neck on the HIII 6-year-old dummy provides improved biofidelity over 
the current dummy and is suitable for compliance purposes, this final 
rule adopts the dummy into FMVSS No. 213 as proposed.
    ii. Weighted 6-Year-Old Dummy. A majority of commenters raised 
concerns with the biofidelity of the weighted 6-year-old-dummy, which 
is intended to model a 50th percentile 8-year-old child. IIHS and NTSB 
commented on the importance of height in measuring

[[Page 37645]]

seat belt fit and injury criteria, particularly head excursion. Both 
determined that the weighted dummy failed to accurately represent the 
height of booster occupants. NTSB stated that the addition of weight to 
the dummy's spine and pelvis was not representative of weight 
distribution in an actual child. Ford expressed concern that the 
weighting of the 6-year-old dummy could result in inaccurate output of 
the injury criteria. Ford expected the weighted dummy to show 
abnormally high chest deflection and abnormally low chest acceleration, 
and higher head excursion. Ford was also concerned that the low 
relative mass of the lower extremities could reduce knee excursion 
compared to a more biofidelic dummy. Ford stated that adding mass to 
the spine and lengthening the lumbar spine might result in the weighted 
dummy not submarining under conditions that would cause a more 
biofidelic dummy to submarine. Public Citizen, Graco, and the Alliance 
commented that the weighted dummy would not perform the same as the 10-
year-old dummy which NHTSA has been developing and which was referenced 
in Public Law 107-318 (Dec. 4, 2002; 116 Stat. 2772)(``Anton's 
Law'').\12\
---------------------------------------------------------------------------

    \12\ On December 4, 2002, Congress enacted Public Law 107-318 
(Anton's Law) ``to provide for the improvement of the safety of 
child restraints in passenger motor vehicles, and for other 
purposes.'' Section 4 of Public Law 107-318 directed that--
    (a) Not later than 24 months after the date of the enactment of 
this Act, the Secretary shall develop and evaluate an 
anthropomorphic test device that simulates a 10-year old child for 
use in testing child restraints used in passenger motor vehicles.
    (b) Within 1 year following the development and evaluation 
carried out under subsection (a), the Secretary shall initiate a 
rulemaking proceeding for the adoption of an anthropomorphic test 
device as developed under subsection (a).
    Other provisions relating to child restraint performance were 
also included in the statute.
---------------------------------------------------------------------------

    IIHS, ACTS, Public Citizen, the Alliance, and GM stated that the 
lack of biofidelity should preclude the use of the weighted dummy. Many 
commenters urged the agency to develop the 10-year-old dummy as an 
alternative. Public Citizen urged the agency to move ahead with 
regulations in anticipation of the 10-year-old dummy's future 
availability. NTSB suggested using the European 10-year-old dummy (P-
series) as an interim measure. While acknowledging the existence of 
problems with the P-series, NTSB stated that European dummy would 
better represent height and seat belt fit.
    While raising concerns with biofidelity, a number of commenters 
agreed that, if necessary, the weighted 6-year-old dummy could be used 
in a limited capacity to test the structural integrity of child 
restraints until such time as the Hybrid III 10-year-old dummy became 
available. Evenflo also supported using the weighted dummy to measure 
head excursion.
    The agency agrees that the Hybrid III 10-year-old dummy, envisioned 
by Anton's Law, represents the long-term solution to the issue of 
testing booster seats certified for higher weights. Development of the 
Hybrid III 10-year-old dummy is proceeding as quickly as possible, but 
this dummy is not currently ready for use in compliance tests. The 
agency is currently testing the Hybrid III 10-year-old dummy to 
determine its suitability for FMVSS No. 213 compliance testing. A 
notice proposing to incorporate this dummy into Part 572 for use in 
compliance testing is expected to be published in early 2004.
    Despite limited results showing a general correlation between the 
testing performance of the weighted 6-year-old dummy and the Hybrid III 
6-year-old dummy, the agency is persuaded by the comments that the 
weighted dummy should not be used for testing with full 
instrumentation. The weighted dummy would not perform the same as the 
10-year-old dummy in development and it may not accurately represent an 
8-year-old child. IIHS stated that the weighted dummy is too short to 
represent the tallest occupants for whom boosters are recommended, 
noting that ``[s]itting height is an important factor in testing 
booster seats because a poorly designed booster may permit too much 
head excursion for taller occupants. Weight is, at most, a secondary 
issue for the restraints because the vehicle belts, which are not 
subject to testing under this standard, restrain the inertia of booster 
seat occupants.''
    While the 0.7-inch increase in sitting height achieved through the 
addition of weights to the Hybrid III 6-year-old dummy is comparable to 
that of a 50th percentile 8-year-old child, the overall weight and 
height, and consequently the weight distribution, are not. The 50th 
percentile 8-year-old child is 50.5 inches tall, as compared to the 
50th percentile 6-year-old child which is 45.5 inches tall. The weight 
added to the 6-year-old dummy is not distributed as it would normally 
be on a 50th percentile 8-year-old, making injury measurements suspect.
    The agency agrees that the kinematics of the weighted 6-year-old 
dummy may not be representative of the older child that it attempts to 
model and it could potentially interact with the belt system 
differently than a dummy developed to represent an 8-year-old child. 
Therefore, the weighted dummy will be used only as a means of ballast 
to evaluate the structural integrity of the tested child restraint. 
While the weighted dummy will not be instrumented to determine 
compliance, it will be instrumented to collect data for use in 
research.
    Anton's Law \13\ directs the agency to initiate a rulemaking 
proceeding to establish performance requirements for child restraints, 
including booster seats, for the restraint of children weighing more 
than 50 lb. Through use of the weighted 6-year-old dummy, the 
structural integrity of a CRS recommended for children between 50 and 
65 lb can be tested. NHTSA recommends children to be placed in booster 
seats until they are 8-years old, or 57 inches tall. The weight of a 
50th percentile 8-year-old male is approximately 57 lb. The weight of a 
50th percentile 8-year, 9 month-old male is approximately 62 lb. Use of 
the 62 lb weighted dummy as ballast ensures that booster seats 
certified up to 65 lb will not structurally fail in a crash.
---------------------------------------------------------------------------

    \13\ Section 3 of Public Law 107-318 directs the Secretary of 
Transportation to consider whether to include injury performance 
criteria for child restraints, including booster seats and other 
products for use in passenger motor vehicles for the restraint of 
children weighing more than 50 pounds.
---------------------------------------------------------------------------

    While several commenters suggested using alternative dummies as an 
interim measure, none of the suggested alternatives are appropriate 
even for use as ballast. NTSB recommended using the European P-series 
10-year-old dummy in a limited capacity to provide a better means of 
evaluating proper seat belt fit and to enhance efforts to enact booster 
seat laws in the states. NHTSA is not confident in the ability of the 
P-series dummy to uniformly load the restraint system in a manner 
necessary for the evaluation of the booster seat, even structurally. 
The P-series dummy is designed with too many degrees of freedom, and 
its interaction with a restraint system would be inconsistent.
    AAP suggested using the Hybrid III 5th percentile female to test 
child restraints to allow regulation up to 80 lb in advance of the 
availability of the Hybrid III 10-year-old dummy. The weight of the 
Hybrid III 5th percentile female dummy is 108 lb, 28 lb heavier than 
the maximum weight of a child that the child restraint would be 
certified for in compliance testing. The heavier weight of the 5th 
percentile female dummy would not offer an accurate representation of 
an 8-year-old or even 10-year-old child.
3. Specific Issues Relating to the Use of the New Dummies in Standard 
No. 213
    i. Seat Back Height Requirement. S5.2.1.1 specifies that each child

[[Page 37646]]

restraint system shall provide head restraint by means of a continuous 
seat back. Subsection (a) of S5.2.1.1 specifies that for child 
restraints recommended for use by children weighing less than 20 lb, 
the height of the seat back must be not less than 18 inches. If a 
restraint were recommended for children weighing 20 to 40 lb, the seat 
back height must be not less than 20 inches.
    Some rear-facing infant car seat/carriers, which are designed with 
a handle for toting the infant outside of the vehicle, are recommended 
for use with infants weighing only up to 20 lb. Under current S5.2.1.1, 
these restraints (recommended for children up to 20 lb) must have a 
seat back of a height of not less than 18 inches. This final rule 
amends S5.2.1.1 to require these restraints to have a seat back height 
of not less than 20 inches.
    The agency proposed to use the CRABI dummy in place of the 9-month-
old dummy in all tests in which the latter dummy is used, including 
tests of rear-facing infant car seat/carriers. Thus, it was proposed 
that the CRABI (at 22 lb) would be used to test car seat/carriers. 
Comments were requested on the appropriateness of using the CRABI dummy 
to test infant car seat/carriers recommended for children up to 20 lb, 
when the 22-lb dummy is heavier than the children recommended for the 
restraints. Comments were requested on whether all infant car seat/
carriers have back supports that are high enough to support the CRABI.
    No commenter opposed the use of the CRABI in place of the 9-month 
old dummy, but some issues were raised about possible effects of using 
the dummy to test infant seats. Graco suggested that S5.2.1.1 could be 
deleted, for lack of a safety need, if Standard No. 213 were amended to 
specify use of the CRABI dummy to assess the ability of a rear-facing 
restraint to limit the rearward excursion of the dummy in Standard No. 
213's dynamic test (S5.1.3.2).\14\ Evenflo stated that several infant-
only restraints do not have backs high enough to support the CRABI 12-
month-old dummy. The commenter suggested that replacement of the 9-
month-old dummy by the CRABI in 4 years would help minimize the 
financial impact to child restraint manufacturers.
---------------------------------------------------------------------------

    \14\ Under S7.1(c) of Standard No. 213, child restraints 
recommended for use by children weighing 22 to 27 lb are tested with 
the 3-year-old (33 lb) dummy. Graco suggested that a weighted CRABI 
12-month-old or an 18-month-old dummy be used instead of the 3-year-
old dummy. Given the agency's resources and the safety issues before 
the agency, NHTSA will not be undertaking rulemaking at this time on 
the weighted CRABI or on an 18-month-old dummy.
---------------------------------------------------------------------------

    In response to Graco, NHTSA agrees that S5.2.1.1 and S5.1.3.2 both 
provide protection to a rear-facing child in a frontal impact by 
limiting occupant excursion outside of the confines of the restraint 
system. However, the agency is unable to conclude that the two 
requirements serve the same safety need for rear-facing restraints. 
S5.2.1.1 specifies seat back height and width requirements and also 
limits how far rearward the test dummy's head may rotate during dynamic 
testing. These requirements may provide protection in dynamic 
conditions other than that replicated by the Standard No. 213 sled 
test. A child restraint might be able to meet S5.1.3.2 with a seat back 
that is lower or narrower than that specified by S5.2.1.1. Deleting 
S5.2.1.1's requirements for rear-facing restraints could reduce some of 
the current protections afforded by child restraints. Thus, the agency 
declines to delete S5.2.1.1.
    At the same time, however, the agency has concluded that with the 
incorporation of the CRABI dummy into the standard, amendments to 
S5.2.1.1 are in order. Information indicates that infants should be 
positioned rear-facing until at least 12-months old, until such time 
their neck and muscular structure are developed to more adequately 
support their head. If rear-facing infant seats were recommended for 
use with an infant until the infant weighs 22 lb, there is a greater 
likelihood that parents will keep their infants in the rear-facing 
restraint until the infant reaches or is closer to reaching 12 months 
of age than if the restraint were only recommended for infants up to 20 
lb. (The agency believes that many infants are positioned forward-
facing in a toddler restraint after being transitioned out of a rear-
facing car seat/carrier, and that many of these infants are not 
developmentally ready to be forward-facing in the vehicle.)
    The agency is amending S5.2.1.1(a) to encourage the production of 
rear-facing infant car seat/carriers that are recommended for use by 
infants up to 12 months in age. The agency is amending the table in 
S5.2.1.1(a) such that infant car seat/carriers must have a minimum seat 
back height of 20 inches.\15\ The effect of this is to require all 
rear-facing infant restraints to be large enough for an average 12-
month-old. As a practical matter, this is not a drastic change. 
Seventy-five percent of the infant-only seats that have been evaluated 
in the agency's ease-of-use ratings program were certified for children 
weighing up to 22 lb and thus already are manufactured with 20-inch 
seat backs.
---------------------------------------------------------------------------

    \15\ More specifically, the section is amended to specify that 
restraints certified for children weighing less than 40 lb must have 
a minimum seat back height of 20 inches.
---------------------------------------------------------------------------

    This final rule does not require manufacturers to recommend on the 
labels accompanying infant restraints that the restraints are 
recommended for infants up to 22 lb, but provides the incentive for 
them to do so. Because the 22-lb CRABI will be the test instrument used 
in compliance tests of the infant seats, and because under S5.2.1.1(a) 
the infant seats must have a minimum seat back of 20 inches, the agency 
believes that manufacturers will certify most if not all infant 
restraints to 22 lb.
    The agency is providing for a 2-year leadtime for this change. 
Evenflo stated that several models of infant-only restraints do not 
have backs high enough to support the CRABI 12-month-old dummy and will 
thus have to be redesigned. Evenflo suggested that replacement of the 
9-month-old dummy by the CRABI in 4 years would help minimize the 
financial impact to child restraint manufacturers. JPMA suggested a 3 
year leadtime. NHTSA declines to provide such long leadtimes suggested 
by Evenflo and JPMA because there could be safety benefits associated 
with keeping more infants rear-facing until they are at least 12-months 
old, which could result from the change to the CRABI and to S5.2.1.1 of 
Standard No. 213. The short deadlines of the TREAD Act also indicate 
Congress's interest in having the standard be upgraded as quickly as 
possible. The 2-year leadtime NHTSA is providing balances the safety 
benefits with the need for some child restraint manufacturers to modify 
some of their seats.
    ii. Padding Requirement. The agency asked for comment on deleting 
S5.2.3, which specifies a padding requirement for child restraints used 
by children weighing less than 22 lb. The agency had specified the 
requirement (whose thickness and static compression specifications are 
compliance-tested statically) because there was no instrumented infant 
test dummy available at the time (1979) the requirement was adopted. 
The agency's goal was to establish dynamic test requirements for infant 
restraints, so that the total energy absorption capability of the 
padding and underlying structure could be measured. (44 FR 72131, 
72135). Graco and Xportation supported deleting S5.2.3. Since today's 
final rule incorporates use of the instrumented CRABI 12-month-old 
dummy for use in testing restraints recommended for children under 22 
lb, we are deleting S5.2.3, as proposed.

[[Page 37647]]

4. Leadtime
    The agency proposed in the NPRM that manufacturers be provided two 
years of leadtime, after publication of a final rule, before specifying 
the use of the CRABI and Hybrid III dummies in compliance tests. The 
NPRM proposed using the weighted 6-year-old dummy in compliance tests 
180 days after publication of a final rule. JPMA supported the addition 
of the new dummies to the standard, provided that the agency gives ``a 
phase in of at least three years from the issuance of the final rule * 
* * to avoid costly recertification requirements for existing seats, 
and to avoid the possible elimination of some current seats from the 
marketplace.'' JPMA stated that because of dimensional differences 
between the proposed CRABI 12-month-old and the 9-month-old dummy 
currently used to test infant-only child restraints, the commenter 
believed that the use of the CRABI dummy will likely result in the 
elimination of current infant-only child restraints. JPMA stated that 
``millions of dollars of tooling and development testing will be 
rendered worthless'' by incorporating the new dummies and that 
``[m]anufacturers should be given a longer lead time before having to 
endure the several financial consequences of these changes.'' \16\ 
Evenflo commented that the agency ``must recognize that the use of the 
new dummies will have a significant affect [sic] on manufacturers' test 
costs, which will ultimately be reflected in the price of child 
restraints.''
---------------------------------------------------------------------------

    \16\ The commenter supported the proposal in the NPRM of 
allowing manufacturers the option of using the new dummies before 
the mandatory compliance date of the standard.
---------------------------------------------------------------------------

    The agency is providing for a 2-year leadtime for the changeover to 
the new dummies. As explained above, the agency believes there are 
safety benefits associated with keeping more infants rear-facing until 
they are at least 12-months old, which could result from the change to 
the CRABI and to S5.2.1.1 of Standard No. 213. At the same time, the 
two year leadtime is provided to lessen the cost impacts of the rule on 
manufacturers' testing costs (retesting current child restraints on the 
new seat assembly using the new dummies, and at test speeds closer to 
30 mph) and possible retooling costs.
    NHTSA believes there also are safety benefits to testing the 
structural integrity of child restraints recommended for children 
weighing from 50 to 65 lb. However, an effective date short of 
approximately two years is not provided for use of the weighted dummy 
because the rulemaking incorporating the dummy into 49 CFR part 572 is 
not yet completed. The NPRM was published May 7, 2003; 68 FR 24417. The 
rulemaking should be completed with sufficient time to allow 
manufacturers to certify their restraints to Standard No. 213 by the 
two-year compliance date.

d. Application of the Standard

    Most commenters supported increasing the weight limit in the 
definition of ``child restraint system'' above the current 50 lb. The 
only commenter opposed to any increase was the Automotive Coalition for 
Traffic Safety, because of concern with the weighted 6-year-old dummy. 
Of those supporting an increase, a majority supported increasing the 
weight to 65 lb based on the use of the weighted 6-year-old dummy, with 
future amendments increasing the weight to 80 lb upon the introduction 
of the 10-year-old dummy. Advocates stated that it would support 
increasing the limit to 65 lb upon showing that the weighted 6-year-old 
dummy (62 lb) is sufficient to assess child restraint use with children 
weighing up to 65 lb. Graco suggested that the agency should defer 
increasing the limit to the time the 10-year-old dummy is available.
    Several commenters did not support an intermediate level of 65 lb 
and preferred amending the standard now to specify the application to 
restraints recommended for children up to 80 lb. Ms. Bidez supported 
incorporating the 10-year-old dummy in its current form and amending 
the weight limit to reflect the 80 lb weight of the 10-year-old dummy. 
AAP recommended using the 5th percentile female to allow regulation up 
to 80 lb in advance of the 10-year-old dummy. E-Z-On believed that the 
limit should be extended to 80 lb, and that costs to vehicle and child 
restraint manufacturers to provide stronger anchorages and hardware 
would be minimal.
    The agency agrees with commenters in that the weight limit in the 
definition of ``child restraint system'' should be increased above 50 
lb. While the weighted 6-year-old dummy injury measurement reliability 
may not be sufficient for compliance testing, the dummy is suitable for 
testing the structural integrity of child restraints up to 65 lb. Use 
of the weighted dummy provides an interim weight limit in advance of 
the Hybrid III 10-year-old dummy. The agency is confident in the 
ability of the 62-lb-weighted dummy to test restraints certified up to 
65 lb. There will be only a 3-lb difference between the weighted dummy 
and the maximum certification weight. The Hybrid II 3 year-old, 
weighing 33 lb, has proven efficient at testing child restraint systems 
certified with a maximum weight of 40 lb.
    However, the weighted 6-year-old dummy is not sufficient to assess 
the dynamic performance of a child restraint in restraining an 80-lb 
child, and as stated above, use of an alternative dummy to allow 
increasing the limit to 80 lb is not appropriate. The agency is not 
confident in the ability of the European P-series 10-year-old dummy to 
uniformly load the restraint, and the Hybrid III 5th percentile female 
is 35 percent heavier than the suggested maximum weight of 80 lb.
    For the aforementioned reasons, NHTSA is increasing the reference 
to the weight limit in the definition of ``child restraint system'' 
from 50 lb to 65 lb. This amendment, effective in 180 days, affects 
primarily manufacturers of child restraints recommended for older 
children, i.e., booster seat and harness manufacturers. The agency does 
not anticipate that manufacturers will have to redesign their 
restraints to certify compliance using the weighted 6-year-old dummy. 
However, the rulemaking to incorporate the weighted 6-year-old dummy 
into part 572 is not complete, so the agency is specifying that 
compliance testing with the weighted dummy will not begin for two 
years. Manufacturers are permitted the option of voluntarily using the 
weighted dummy prior to the mandatory compliance date.
    Several comments were submitted on whether manufacturers should be 
prohibited from recommending their seats for children of weights higher 
than the heaviest dummy used to test the restraint. Consumer Union 
stated that the agency should limit manufacturers' ability to advertise 
child restraint weight maximums only to the weight of the heaviest 
dummy used for its certification testing. Alternatively , Consumer 
Union stated that the agency should develop dummies that are at the 
maximum weight advertised for the restraint, or require the addition of 
ballast weights to existing test dummies.
    In contrast, TraumaLink believed that manufacturers should be 
permitted to recommend child restraints at weights above that of the 
heaviest dummy used to test the restraint. TraumaLink stated that there 
was no field data to indicate a problem with convertible restraints 
(typically recommended for children up to 40 lb) which have been tested 
with a 33 lb dummy (the Hybrid II 3-year-old). Limiting the regulation 
based on the heaviest dummy, TraumaLink continued, would place 
artificial limits on the protections afforded children. Similarly, AAP 
opposed limiting a

[[Page 37648]]

manufacturer's ability to recommend a child restraint for a weight 
above that of the heaviest dummy used to test the restraint. AAP stated 
that such a restriction could mislead parents into thinking that 
children should use seat belts once the child is heavier than 62 lb, 
when in fact, most children do not fit seat belts until a much heavier 
weight.
    In a rulemaking amending FMVSS No. 213 to incorporate several test 
dummies into the standard (61 FR 30827; June 18, 1996), NHTSA responded 
to Consumer's Union (CU) belief, expressed during that rulemaking, that 
restraints (e.g., convertible child seats) should not be permitted to 
be recommended for children weighing more than the largest test dummy 
used to test the restraint (e.g., 33 lb). The agency determined that 
such an approach was unnecessarily restrictive, given that there has 
been no showing that the wider array of dummies incorporated into 
Standard No. 213 by that rulemaking were insufficient surrogates for 
the children for whom the restraints are recommended. The agency also 
believed that CU's suggestion could have unintended safety 
consequences, because it would have the effect of forcing young 
children out of child restraints specifically designed for them 
(typically 20 to 40 lb) and into restraints that may not be appropriate 
for their size, i.e., booster seats for a 3-year-old or the vehicle's 
belt systems. The agency believed that while it might be hypothetically 
possible that a restraint that passed FMVSS No. 213 when tested with a 
dummy could fail when restraining a child weighing slightly more than 
the dummy, on balance, the possibility of such a failure is outweighed 
by the safety risk of forcing children into restraints that might not 
adequately restrain them.
    NHTSA reaffirms the conclusions reached in that rulemaking and 
concurs with the views of TraumaLink and AAP that information on tests 
with current test dummies does not indicate a need to restrict 
recommending child restraints for children weighing more than the test 
dummies used to test the restraint. As to CU's suggestion for 
developing dummies that reach the maximum weight recommended for a 
restraint or requiring the addition of ballast weights to existing 
dummies, this suggestion is beyond the scope of the present rulemaking.

e. Injury Criteria

1. Post-NPRM Testing
    i. JPMA. In its comment to the NPRM, JPMA stated that it had 
conducted a series of 80 sled tests at Veridian Engineering in response 
to the proposal, to try to understand how the proposed dummies 
performed compared to the dummies currently in use. The tests also 
evaluated the proposed changes to the standard bench seat, as well as 
the proposed injury criteria. JPMA described its test plan as including 
all test modes for all of the proposed dummies with representative 
samples of all types of child restraint/harness combinations and 
installation methods, including lap belt only, lap/shoulder belt, and 
LATCH. JPMA acknowledged that: ``While a total of 80 tests were 
conducted, this series only begins to explore the results of the 
proposed changes and does not allow analysis of the net effect of each 
change, nor does it provide enough history to define the potential 
variability in test results which could occur. Much more testing is 
required to define the new effect of each change and the potential 
variation which can have a significant impact on design and ability to 
define compliance margins.''
    ii. NHTSA Series I and II. PAX conducted a series of dynamic sled 
tests for NHTSA to evaluate the performance of various child restraints 
on the revised test seat assembly. The tests used the CRABI and Hybrid 
III 3- and 6-year-old dummies to evaluate whether these dummies could 
meet the proposed scaled HIC, chest injury limits and Nij measures. 
Time and resource considerations limited the testing to 5-point harness 
rear-facing infant seats, convertible safety seats, and belt-
positioning seats. Restraints were evaluated while installed using a 
lap belt, a lap/shoulder belt, and the LATCH system. HIC measurements 
were obtained, but testing problems arose with respect to the neck 
injury and chest deflection data. Because of these problems, NHTSA 
conducted a second series of dynamic sled tests at VRTC to replicate 
the tests performed at PAX with the 3- and 6-year-old dummies. These 
tests were conducted using the same type of child restraints.
    iii. Results of JPMA and NHTSA Series I and II. The charts provided 
in a docket submission titled ``Comparison of PAX/VRTC and JPMA Sled 
Tests'' summarizes the results of the testing performed by both NHTSA 
and by JPMA. For HIC and chest acceleration, results are presented for 
both the proposed scaled injury limits and for the same tests using the 
existing injury criteria limits stated in FMVSS No. 213. Because chest 
deflection and Nij limits are not currently specified in FMVSS No. 213, 
the only charts provided are for the proposed criteria limits.
    NHTSA testing performed at PAX and VRTC, described in the NPRM, 
resulted in dummy responses that were generally within the injury 
limits proposed in the NPRM, with the exception of Nij. (While 
acceptable Nij readings were found in tests using the Hybrid III 3-
year-old dummy, there were widespread failures in both rear- and 
forward-facing tests using the CRABI 12-month old dummy and each of 
four tests with the Hybrid III 6-year-old dummy.) However, the test 
results presented by JPMA were quite different. JPMA's testing, using 
the revised test seat assembly and new dummies as NHTSA tested, but 
across a wider variety of child restraint types, showed very mixed 
results. In many instances, the measured injury parameters either 
exceeded or marginally passed the scaled injury limits proposed in the 
NPRM. Further, there were many JPMA tests that resulted in either 
failing or marginal results when using the existing injury criteria. 
This raised questions regarding the combined effect of the changes to 
the test seat assembly, incorporation of the new dummies, and use of 
the scaled injury criteria limits together.
    iv. NHTSA Series III. In an effort to determine if the use of 
varying restraint types in the JPMA testing (as opposed to NHTSA's use 
of 5-point harness restraints only) could be identified as the 
predominant factor in explaining the disparity between the JPMA and 
NHTSA test results, NHTSA conducted a third series of sled tests. These 
tests were performed at VRTC, and attempted to closely parallel the 
testing performed by JPMA. In addition to a number of additional 5-
point harness restraints, NHTSA also tested forward-facing convertible 
overhead shield child restraints, and shield-type boosters both with 
and without the shield.
    A total of 20 additional tests were conducted in this third series 
of sled tests. The results of this series of sled tests more closely 
paralleled those found in the JPMA tests, in that a wider range of 
failing and marginal test results were seen as compared to the 
predominately passing results seen in the PAX test series. The testing 
of 5-point harness restraints at PAX resulted in injury values that 
were predominately within the established or proposed limits (with the 
exception of Nij). However, the VRTC Series III tests showed a wider 
variation in both marginal and failing responses that appear to be a 
result of the restraint type

[[Page 37649]]

that was tested, even though all restraint types meet the current FMVSS 
No. 213 requirements and appear to be equally effective based upon 
field studies. Not all VRTC results were similar to those of JPMA, 
however, as the HIC15 results for the CRABI dummy were typically better 
in the VRTC tests than in the JPMA tests. Overall, the VRTC follow-on 
tests did confirm the wider range of test results found in the JPMA 
testing. The follow-on testing results can be found in the docket.
2. Comments and Conclusions
    i. Head Injury Criterion (HIC). The agency received widely 
divergent comments on the proposal to limit measurement of HIC to 15 
milliseconds and to use the injury criteria of Standard No. 208 that 
were scaled for children. The Alliance, UMTRI and SafetyBeltSafe 
supported the use of a 15 ms limit on the head injury criterion (HIC) 
limit as a more realistic way to assess head and brain injury, with the 
lower HIC values proposed for each dummy. In contrast, Advocates stated 
that it was ``reluctant to change the duration of HIC measurement from 
36 ms to 15 ms without more definitive evidence that this change would 
not inhibit accurate HIC measurements associated with non-contact head 
injuries.'' Advocates suggested that NHTSA should scale the injury 
assessment reference values for children even if the agency decides not 
to shorten the HIC measurement duration, to ``take into account the 
different anatomy of children.'' Ford stated that, while the purpose of 
the 15 ms limit on the HIC calculation interval is to try to 
differentiate between HICs caused by hard head contacts and non-contact 
HICs due to head whipping, the 15 ms HIC measurement should not be used 
to differentiate between non-contact and ``chance contact'' of the 
dummy's head with the dummy's knees.
    JPMA stated that it was willing to consider supporting a 15 ms 
limit (HIC15), if the agency can undertake research to 
assure that there will not be unintended consequences from 
countermeasures needed to meet HIC15. However, JPMA did not 
support the other proposed new injury criteria, including the scaled 
HIC values. The commenter stated that the tests of child restraints it 
conducted with the proposed CRABI and Hybrid III dummies produced 
injury reference values that exceeded the proposed limits, which the 
commenter said is a concern given the high level of effectiveness of 
current child restraints. The commenter suggested that it might be more 
feasible to use the FMVSS No. 208 criteria in FMVSS No. 213 if the 
agency were to specify a ``more realistic crash pulse for FMVSS No. 
213, such as the one contained in the FMVSS No. 208 sled test.'' Graco 
questioned why the scaled HIC values would be applied to in-position 
child restraint testing if they were derived from out-of-position 
occupant airbag testing. Graco believed that the values might not be 
``applicable to child restraint testing with a 213-style pulse.'' The 
commenter stated that it saw minimal benefit to child passenger safety 
from using the proposed injury criteria. It was concerned that some 
seats that have historically performed well in the real world and in 
compliance testing would fail the new criteria.
    Response: This final rule retains the existing FMVSS No. 213 HIC 
threshold of 1000 for the CRABI 12-month-old and Hybrid III 3- and 6-
year-old dummies.
    Since the TREAD Act directed NHTSA to consider adopting the scaled 
injury criteria adopted by the May 2000 final rule on advanced air 
bags, NHTSA proposed that the HIC limits of 39015, 
57015 and 70015 be incorporated into FMVSS No. 
213 for tests with the CRABI 12-month-, and Hybrid III 3- and 6-year-
old dummies, respectively. However, NHTSA believed that it should take 
a cautious approach in modifying the head injury tolerance level set by 
the HIC requirement. The agency requested comments on issues related to 
the proposed injury criteria, such as on what risk levels are 
acceptable, what factors should be considered in selecting performance 
limits and whether the same limits as in FMVSS No. 208 should be 
established for the child restraint standard. The agency noted that the 
two standards address different sources of potential harm to children. 
The injury criteria for children in FMVSS No. 208 are intended to 
minimize the risk from a deploying air bag (ensuring that the air bag 
deploys in a manner much less likely to cause serious or fatal injury 
to out-of-position occupants). The injury criteria in FMVSS No. 213 are 
intended to limit the severity of forces imposed on a child during a 
crash. Child restraints meeting these criteria have worked effectively 
to maintain high levels of performance in crashes. Because the injury 
criteria of the standards are intended to minimize risks from different 
injury sources, the agency stated that it might be reasonable to have 
non-identical criteria.
    In this final rule, NHTSA has decided against incorporating the 
scaled injury limits used in FMVSS No. 208 because the data obtained 
from the JPMA and NHTSA (series III) test programs indicate that 
current child restraints generally do not meet the proposed limits. 
There are several reasons why this was a concern for the agency. First 
and foremost, child restraints are currently highly effective in 
reducing the likelihood of death and/or serious injury in motor vehicle 
crashes. The agency was unable to identify a safety problem that the 
scaled injury limits of FMVSS No. 208 would remedy.
    Second, it is unknown what modifications to child restraints would 
be necessary for the restraints to meet the proposed injury limits. 
Commenters did not provide information on how child restraints that 
failed to meet the proposed Nij and other limits could be modified to 
meet the criteria. Assuming that the restraints could be redesigned to 
meet the proposed injury limits, there would likely be costs associated 
with the redesign which would result in increases in the price of the 
restraints. As noted above in section IV of this preamble, the agency 
considers the consumer acceptance of cost increases to child restraints 
(an already highly-effective item of safety equipment) in determining 
the net safety effects of changes to the child restraint standard. In 
balancing the effects of meeting the scaled injury criteria against the 
possible impacts on the price of restraints, the agency determined that 
the net effect on safety could be negative in this instance because of 
the minimal benefits of such a change, weighed against the delayed 
replacement of old restraints by current owners or non-purchase by non-
owners. For these reasons, in accordance with the TREAD Act, we have 
considered whether to apply scaled injury criteria performance levels 
developed for FMVSS No. 208 to child restraints and have determined it 
would not be prudent to do so.
    NHTSA is adopting HIC36 with a limit of 1000 for all 
tests with the Hybrid III and CRABI dummies. This final rule does not 
adopt the 15 ms window that was proposed in the NPRM. This is because 
the shorter time interval would likely substantially reduce the values 
calculated for the HIC in compliance tests. Further, as discussed later 
in this section, NHTSA is not incorporating a neck injury criterion 
into FMVSS No. 213. A 36 ms time interval to measure HIC allows the HIC 
measurement in FMVSS No. 213 to capture risk of neck injury indirectly. 
Given that the agency is declining to adopt a neck injury criterion at 
this time, the longer measurement window associated with 
HIC36, as opposed to HIC15, will provide 
reasonable assurances that a child's neck will not be subjected to 
excessive forces in a crash. The 36 ms time

[[Page 37650]]

interval to measure HIC is consistent with the injury threshold used in 
FMVSS No. 208 for the Hybrid III 50th percentile dummy prior to the 
incorporation of scaled injury limits and Nij for advanced air bags.
    Limiting the duration over which HIC is calculated to a maximum of 
36 ms, while limiting HIC to 1000, assures that the acceleration level 
of the child's head will not exceed 60 g's for any period greater than 
36 ms. The 60 g acceleration limit was set as a reasonable head injury 
threshold by the originators of the ``Wayne State Tolerance Curve'', 
which was used in the development of the HIC calculation.
    The change to a 36 millisecond time measurement for HIC will not 
necessarily result in lower HIC values in compliance testing because of 
the changeover in this rulemaking to the new dummy family. NHTSA 
compared the differences between using the HIC36 criterion 
in testing with the Hybrid III dummy family and using the existing 
criterion, HICunlimited, in testing with the Hybrid II 
family. The following tables outline the results of comparison tests 
performed on identical child restraints, using the FMVSS No. 213 
proposed (Table 8) and existing seat assemblies (Table 9), with both 
Hybrid III and Hybrid II 3-year-old dummies.

       Table 8.--Comparison Tests of 3-Year-Old Hybrid III and Hybrid II Dummies on Proposed Seat Assembly
  [Hybrid III HIC36 vs Hybrid II HICUnlimited 3-Year-Old Child Dummy (Tested Using with The NPRM Proposed Seat
                                                   Assembly)]
----------------------------------------------------------------------------------------------------------------
                                                Hybrid III*   Hybrid III
                                                   HIC36     HICUnlimited                  Trend
----------------------------------------------------------------------------------------------------------------
Cosco Touriva Convertible, Lap Belt, No           [sim]434           703   Hybird III HIC36 Less than Hybrid II
 Tether.                                                                    HICUnlimited.
Century Accel Convertible, Lap Belt, No           [sim]344           627   Hybird III HIC36 Less than Hybrid II
 Tether.                                                                    HICUnlimited.
Century Breverra Hybrid, Lap Belt, No Tether.     [sim]521           670   Hybird III HIC36 Less than Hybrid II
                                                                            HICUnlimited.
Cosco HB Booster Hybrid, Lap Belt, No Tether.     [sim]684           446   Hybird III HIC36 Greater than Hybrid
                                                                            II HICUnlimited.
----------------------------------------------------------------------------------------------------------------
* HIC36 were not calculated, the relationship HIC36 = 0.97
* HICUnlimited was used to approximate HIC36.


       Table 9.--Comparison Tests of 3-Year-Old Hybrid III and Hybrid II Dummies on Existing Seat Assembly
  [Hybrid III HIC36 vs Hybrid II HICUnlimited 3-Year-Old Child Dummy (Tested Using Existing FMVSS No. 213 Seat
                                                   Assembly)]
----------------------------------------------------------------------------------------------------------------
                                                Hybrid III     Hybrid II
                                                   HIC36     HICUnlimited                  Trend
----------------------------------------------------------------------------------------------------------------
FF Convertible, Lap Belt.....................          671           385
FF Convertible, Lap Belt.....................  ............          479
FF Convertible, Lap Belt.....................  ............          424
    Average..................................          671           429   Hybird III HIC36 Greater than Hybrid
                                                                            II HICUnlimited.
FF Convertible, Lap and Tether...............          303           387
FF Convertible, Lap and Tether...............          362           396
    Average..................................          333           392   Hybird III HIC36 Less than Hybrid II
                                                                            HICUnlimited.
FF Convertible, LATCH........................          292           281
FF Convertible, LATCH........................          518           336
    Average..................................          408           309   Hybird III HIC36 Greater than Hybrid
                                                                            II HICUnlimited.
FF Hybrid, Lap and Tether....................          452           392
FF Hybrid, Lap and Tether....................          439           501
    Average..................................          446           447   =
----------------------------------------------------------------------------------------------------------------

    In some of the tests Hybrid III HIC36 results were 
higher, and in other tests the HII HICUnlimited results were 
higher. On the other hand, in a limited number of tests with the 6-
year-old dummies, the Hybrid III HIC36 numbers were higher. 
All in all, the agency determined that the data are inconclusive as to 
any differences in how the Hybrid II and Hybrid III dummies measure 
HIC. In any event, the agency's tests of child restraints has not found 
any difference between HICUnlimited and HIC36 in 
terms of compliance passage rates. Thus, the agency has concluded that 
the impact on child restraint performance relating to the change to 
HIC36 will be insignificant.
    ii. Chest Injury Criteria. Some commenters supported while others 
opposed the proposals to adopt a new chest deflection criterion and to 
adopt the chest acceleration limits that were scaled for children and 
incorporated into FMVSS No. 208. The Alliance supported the proposals. 
Ms. Bidez supported the proposed chest deflection criteria, stating 
that ``although no significant reports of chest injury in children have 
yet occurred, prudence and systems engineering dictates [sic] that 
excessive chest deflection be monitored to prevent the introduction of 
``new'' injury mechanisms in the quest to prevent other injury 
mechanisms with improved restraint design.''
    JPMA opposed the proposed chest injury criteria for the reasons 
explained in the preceding section. TraumaLink also opposed 
incorporation of the proposed chest deflection and reduced chest 
acceleration limits, because according to data it has collected in its 
study, ``These types of injuries do not occur in children in [child 
restraint systems].'' TraumaLink further stated: ``We are concerned 
about the tradeoff between including these more restrictive thoracic 
criteria and reducing the overall protection of the head through 
increased head excursions and accelerations.'' These concerns were 
echoed by UMTRI, which stated that the relationship between the chest 
deceleration and deflection limits and field injuries under the type of 
loading simulated in FMVSS No. 213 are not well established. 
``Introducing these injury criteria now [including neck injury] could 
lead to counterproductive child restraint designs because many 
restraints that perform well in the field, particularly booster seats, 
are likely to exceed the new injury tolerance measures.''
    SafetyBeltSafe also opposed the proposed chest injury criteria. It 
expressed concern that the new seat bench assembly has an added slope 
to

[[Page 37651]]

the seat cushion that results in a ``harder stop as the restraint 
bottom[s] out against the plywood platform.'' The commenter was 
concerned that, if the chest acceleration limit were reduced, child 
restraints that are already close to the current limit could fail the 
test with no change in how they actually perform in the field. ``To 
counteract this possibility [of failing the test], a manufacturer could 
soften the system, allowing more head excursion (due again to the 
geometry change), to keep the chest acceleration in check. This would 
obviously be counterproductive to child safety.'' In addition, 
SafetyBeltSafe believed that the proposed chest deflection limit ``does 
not relate to any evident injury among restrained child passengers'' 
and thus would not advance child safety. JPMA, UMTRI and SafetyBeltSafe 
suggested that the agency collect data on chest deflection to establish 
a database that could be used to evaluate these measures more in the 
future.
    Ford stated that in its sled tests of booster seats using the 
Hybrid III six-year-old dummy and the FMVSS No. 213 sled pulse, none of 
the tested boosters could be certified as meeting the proposed limits. 
``Boosters that showed good shoulder belt fit routinely measured chest 
acceleration at or near the 60 g limit and chest deflection very near 
the 40 mm limit. Dummy chest values were sometimes below the compliance 
limit, but were seldom far enough below the limit to provide a 
reasonable compliance margin.'' The commenter believed that boosters do 
improve child safety when used properly, and that ``if dynamic testing 
of boosters is continued, the test procedure needs a major overhaul to 
effectively differentiate between acceptable and unacceptable 
designs.''
    Response: This final rule does not adopt the proposed chest injury 
criteria relating to acceleration and deflection. A safety need for 
adopting the proposal has not been established. NHTSA is persuaded by 
the commenters that there are not sufficient data that demonstrate that 
children have been seriously injured due to excessive chest 
acceleration or deflection in current restraint designs. Historically, 
the majority of child injuries are to the head as opposed to the chest. 
The agency is concerned about possible negative effects of adopting the 
proposed chest injury criteria on increased head excursion, as noted by 
SafetyBeltSafe. Further, not enough is known about the countermeasures 
that could be employed to meet the proposed criteria. If child 
restraint manufacturers were to redesign their restraints to meet such 
requirements, the agency is concerned about the possibility of those 
revised designs compromising other aspects of the occupant's injury 
protection.
    The data presented by JPMA, and to a lesser degree, the follow-on 
tests conducted at VRTC, show difficulty for current restraints to meet 
the scaled chest criteria, and also show problems for certain restraint 
types to meet the existing requirements with the revised test seat 
assembly and new dummies. Redesigning the restraints to meet the 
requirements, assuming such redesign is practicable, would involve a 
cost increase to manufacturers, which would be passed on to consumers. 
The agency does not believe that the cost increase is justified in this 
instance, and is concerned about the possible effect the cost increase 
could have on the purchase and use of child restraints. For the 
aforementioned reasons, we conclude that it is not in the interest of 
safety to adopt the chest injury criteria developed for FMVSS No. 208 
into FMVSS No. 213.
    iii. Neck. Virtually all parties commenting directly on this aspect 
of the proposal opposed the modified Nij neck criterion (modified from 
the criterion in FMVSS No. 208 in that the limits on axial force were 
excluded). The Alliance stated that it believes that serious neck 
injuries in child restraints are most likely caused by excessive upper 
neck tension, and not by exceeding the proposed Nij criterion. The 
commenter thus suggested the agency should specify neck tension and 
compression limits, as follows, when testing with the CRABI 12-month-, 
the HIII three-year- and the HIII six-year-old dummies, respectively: 
780, 1430 and 1890 N for peak tension; and 960, 1380 and 1820 N for 
peak compression. The Alliance further stated, however, that applying 
these limits while maintaining the current sled pulse is likely to make 
compliance with the requirements impossible or possible only with 
substantial cost increases. The commenter suggested that NHTSA modify 
the crash pulse ``to make it more representative of the current crash 
environment'' instead of eliminating neck tension. Ms. Bidez stated 
that ``Nij does not appear to predict cervical distraction injuries in 
children * * *.'' The commenter suggested that ``the duration of the 
axial distraction load as influenced by the presence or absence of 
adequate torso restraint appears to be a more valid predictor of 
(spinal cord injury without radiographic abnormality) SCIWORA injuries 
among children in the absence of head contact.''
    JPMA, SafetyBeltSafe, UMTRI, TraumaLink and others did not support 
adopting the proposed Nij criterion at this time. SafetyBeltSafe 
believed that neither Nij as proposed nor Nij with a limit on tension 
should be used as a compliance criterion unless these are proven to be 
useful predictors of child neck injury. UMTRI believed that Nij should 
not be incorporated at this time because the relationship between the 
criterion and real-world injuries ``under the type of loading simulated 
by FMVSS 213 are [sic] not well established.'' The Insurance Institute 
for Highway Safety (IIHS) was concerned that studies of real-world 
crashes indicate that neck injuries due to inertial forces appear to be 
rare, yet, the commenter stated, it is not clear how child restraints 
could be better designed to lower neck injury measures. Ford stated 
that, in its sled tests of booster seats, ``Upper neck tensions and 
extension moments above the FMVSS 208 criteria were also routinely 
measured. Every test exceeded at least one of the Nij limits.''
    TraumaLink was concerned about the state of knowledge about 
pediatric neck injury and suggested that not enough was known to 
proceed at this time. The commenter stated that data on the 
biomechanical response of the pediatric neck to trauma are severely 
limited and as a result, the neck of current child dummies may not be 
representative of the real child. The commenter also believed that 
efforts to include pediatric neck tolerance levels in regulatory 
efforts are scientifically premature. TraumaLink further stated:

More research is needed to understand the movement of the child's 
neck in traumatic events and the likelihood for injury before 
enacting regulatory standards, but our results indicate that this 
work is of paramount importance. We believe that this research may 
reveal the importance of neck tension and suggest that exclusion of 
limits on peak tension in the test procedure is not appropriate. 
Therefore, we feel that the neck injury measures should be 
calculated but not used in the pass/fail criteria in the FMVSS 213 
test to build the fund of knowledge needed to further refine the 
injury measure.

    Similarly, commenters JPMA, SafetyBeltSafe, UMTRI and the IIHS 
suggested that more research is needed on neck injury among restrained 
children. Some of these suggested that NHTSA measure neck force and 
moment parameters during compliance tests to become familiar with the 
range of results.
    Response: The agency has decided not to incorporate Nij into FMVSS 
No. 213 compliance tests at this time. Both NHTSA and JPMA testing has 
clearly demonstrated that existing child

[[Page 37652]]

restraints that have historically performed very well in the field 
cannot meet the proposed neck injury limits in the majority of test 
cases. Neither NHTSA nor child restraint manufacturers have identified 
any countermeasures that could be incorporated into existing designs 
that would promote compliance with the proposed requirements. Further, 
NHTSA agrees that there is a lack of injury data to demonstrate a need 
to incorporate neck injury criteria at this time. As discussed in the 
section regarding head injury criterion, the adoption of a 36 ms 
measurement window for HIC, as opposed to the 15 ms window that was 
presented in the NPRM, will also serve as surrogate of sorts for a neck 
injury criterion to ensure that children continue to be well protected.
    NHTSA does not believe that enough is known regarding neck injury 
for children at this time. As the agency is not proposing the 
incorporation of Nij in this final rule, NHTSA likewise does not feel 
that it is appropriate at this time to specify neck tension limits or 
any other neck criterion. These are areas where the agency could 
perform additional research in the coming years, as warranted by a 
safety need and the demands on the agency's resources.
    In accordance with the TREAD Act, NHTSA has considered adopting the 
neck injury criteria developed for FMVSS No. 208 into FMVSS No. 213. 
For the aforementioned reasons, we conclude that incorporating the 
criteria into Standard No. 213 is not warranted at this time.

f. Leadtime

    The TREAD Act required NHTSA to complete this rulemaking by 
November 1, 2002. With that date in mind, the agency made the following 
conclusions about the dates on which compliance with the requirements 
will become mandatory.
    a. NHTSA believes that manufacturers could begin certifying their 
child restraints based on testing done on the new seat assembly and 
pulse in approximately 2 years (i.e., the effective date for the change 
will be August 1, 2005). NPRM proposed a 2-year leadtime, which Graco 
supported. While the agency does not expect the changes to the seat 
assembly to have a major effect on the results of compliance tests, 
restraint manufacturers will likely have to conduct testing to confirm 
compliance of their restraints. This will be a financial impact on the 
manufacturers that could be spread out over a 2-year time period. The 
agency does not anticipate any lives saved or injuries avoided from the 
amendment.
    b. This final rule provides for about a 2-year effective date for 
the requirement to use the new CRABI and Hybrid III dummies in 
compliance tests (the effective date for the change will be August 1, 
2005). The agency does not expect that the changes to the dummies will 
have a significant effect on the results of compliance tests, with the 
exception of some infant-only car seat/carriers. However, restraint 
manufacturers will likely have to conduct testing to confirm compliance 
of their restraints. This will be a financial impact on the 
manufacturers that could be spread out over a 2-year time period. Some 
infant-only restraints do not have backs high enough to support the 
CRABI 12-month-old dummy and will thus have to be redesigned.
    The agency cannot estimate any lives saved or injuries avoided from 
the amendment. There could be safety benefits associated with keeping 
more infants rear-facing until they are at least 12-months old, which 
could result from the change to the CRABI and to having infant car 
seat/carriers be designed with higher back support structures.
    c. As for using the weighted 6-year-old dummy to test restraints 
(typically booster seats) recommended for children with masses of over 
22.7 kg (weights over 50 lb), this rule specifies a 2-year leadtime for 
the requirement (the effective date for the change will be August 1, 
2005). We do not anticipate that manufacturers will have to redesign 
their booster seats or safety harnesses to certify compliance using the 
dummy. However, the rulemaking to incorporate the weighted 6-year-old 
dummy into part 572 is not complete, so the effective date is provided 
to account for the completion of that rulemaking. (The part 572 NPRM 
was published May 7, 2003; 68 FR 24417.)
    d. Manufacturers are permitted the option of voluntarily using the 
new sled assembly and pulse and the new test dummies prior to the date 
(August 1, 2005) on which they would be required to do so. Note, 
however, that this final rule also specifies that a manufacturer's 
selection of a compliance option (e.g., to use the new dummies prior to 
the mandatory compliance date) must be made prior to, or at the time of 
the compliance test and that the selection is irrevocable for that 
child restraint. This provision is needed for NHTSA to efficiently 
carry out its enforcement responsibilities. The agency wants to avoid 
the situation of a manufacturer confronted with an apparent 
noncompliance (based on a compliance test) with the option it has 
selected responding to that noncompliance by maintaining that its 
products comply with a different option for which the agency has not 
conducted a compliance test. To ensure that the agency will not be 
asked to conduct multiple compliance tests, first for one compliance 
option, then for another, this rule requires manufacturers to select 
the option by the time it certifies the child restraint system and 
prohibits them from thereafter selecting a different option for the 
restraint.

IX. Rulemaking Analyses and Notices

a. Executive Order 12866 (Regulatory Planning and Review) and DOT 
Regulatory Policies and Procedures

    The agency has considered the impacts of this final rule under 
Executive Order 12866 and the Department of Transportation's regulatory 
policies and procedures. While the NPRM was reviewed under the 
Executive Order, this document was not reviewed because it is 
considerably narrower than the NPRM and has minimal costs. This 
document was treated as ``not significant'' under the Department of 
Transportation's regulatory policies and procedures.
    The estimated costs for this final rule are discussed in NHTSA's 
final regulatory evaluation (FRE) for this final rule.\17\ There is a 
one-time cost of $1.68 million for manufacturers to purchase the new 
test dummies and $1.39 to $3.44 million to certify existing child 
restraints to the new dummies and test requirements. The annual long-
term costs are estimated to be $31,200 to test new models of booster 
seats (including built-in restraints) with a weighted 6-year-old dummy. 
We believe that use of the new dummies, in itself, would not 
necessitate redesign of child restraints. The new dummies perform 
similarly to the ones presently used in compliance testing.
---------------------------------------------------------------------------

    \17\ NHTSA's final regulatory evaluation (FRE) discusses issues 
relating to the potential costs, benefits and other impacts of this 
regulatory action. The FRE is available in the docket for this rule 
and may be obtained by contacting Docket Management at the address 
or telephone number provided at the beginning of this document. You 
may also read the document via the Internet, by following the 
instructions in the section below entitled, ``Viewing Docket 
Submissions.'' The FRE will be listed in the docket summary.
---------------------------------------------------------------------------

    The agency does not believe that updating the seat assembly and 
revising the crash pulse will affect dummy performance to an extent 
that benefits would accrue from such changes, nor will benefits be 
gained by the change to the dummies. There could be safety benefits 
associated with keeping more

[[Page 37653]]

infants rear-facing until they are at least 12-months old, which could 
result from the change to the CRABI and to having infant car seat/
carriers be designed with higher back support structures. However, the 
agency cannot quantify any lives saved or injuries avoided from the 
amendment.

b. Regulatory Flexibility Act

    The Regulatory Flexibility Act of 1980, as amended, requires 
agencies to evaluate the potential effects of their proposed and final 
rules on small businesses, small organizations and small governmental 
jurisdictions. I hereby certify that this final rule will not have a 
significant economic impact on a substantial number of small entities. 
NHTSA estimates there to be about 13 manufacturers of child restraints, 
four or five of which could be small businesses.
    This rule will not generally increase the testing that NHTSA 
conducts of child restraints, except that booster seats, harnesses and 
other types of child restraints that could be recommended for children 
weighing over 50 lb will be tested with the weighted 6-year-old dummy, 
in addition to the dummies presently used to assess the performance of 
the restraint (generally these are the 3-year-old and the unweighted 6-
year-old dummies). Thus, the certification responsibilities of 
manufacturers will not generally be affected. The agency does not 
believe this final rule will impose a significant economic impact on 
small entities, because these businesses currently must certify their 
products to the dynamic test of Standard No. 213. That is, the products 
of these manufacturers already are subject to dynamic testing using 
child test dummies. The effect of this final rule on most child 
restraints is to subject them to testing with new dummies in place of 
existing ones, and/or an additional dummy. Testing child restraints on 
a new seat assembly is not expected to significantly affect the 
performance of the restraints.

c. Executive Order 13132 (Federalism)

    Executive Order 13132 requires NHTSA to develop an accountable 
process to ensure ``meaningful and timely input by State and local 
officials in the development of regulatory policies that have 
federalism implications.'' ``Policies that have federalism 
implications'' is defined in the Executive Order to include regulations 
that have ``substantial direct effects on the States, on the 
relationship between the national government and the States, or on the 
distribution of power and responsibilities among the various levels of 
government.'' Under Executive Order 13132, the agency may not issue a 
regulation with Federalism implications, that imposes substantial 
direct compliance costs, and that is not required by statute, unless 
the Federal government provides the funds necessary to pay the direct 
compliance costs incurred by State and local governments, the agency 
consults with State and local governments, or the agency consults with 
State and local officials early in the process of developing the 
proposed regulation. NHTSA also may not issue a regulation with 
Federalism implications and that preempts State law unless the agency 
consults with State and local officials early in the process of 
developing the proposed regulation.
    We have analyzed this final rule in accordance with the principles 
and criteria set forth in Executive Order 13132 and have determined 
that this rule does not have sufficient Federal implications to warrant 
consultation with State and local officials or the preparation of a 
Federalism summary impact statement. The rule will not have any 
substantial impact on the States, or on the current Federal-State 
relationship, or on the current distribution of power and 
responsibilities among the various local officials.

d. Unfunded Mandates Reform Act

    Section 202 of the Unfunded Mandates Reform Act of 1995 (UMRA) 
requires Federal agencies to prepare a written assessment of the costs, 
benefits and other effects of proposed or final rules that include a 
Federal mandate likely to result in the expenditure by State, local or 
tribal governments, in the aggregate, or by the private sector, of more 
than $100 million in any one year ($100 million adjusted annually for 
inflation, with base year of 1995). (Adjusting this amount by the 
implicit gross domestic product price deflator for the year 2000 
results in $109 million.) This final rule will not result in costs of 
$109 million or more to either State, local, or tribal governments, in 
the aggregate, or to the private sector. Thus, this final rule is not 
subject to the requirements of sections 202 of the UMRA.

e. National Environmental Policy Act

    NHTSA has analyzed this final rule for the purposes of the National 
Environmental Policy Act. The agency has determined that implementation 
of this action will not have any significant impact on the quality of 
the human environment.

f. Executive Order 12778 (Civil Justice Reform)

    This final rule will not have any retroactive effect. Under 49 
U.S.C. 21403, whenever a Federal motor vehicle safety standard is in 
effect, a State may not adopt or maintain a safety standard applicable 
to the same aspect of performance which is not identical to the Federal 
standard, except to the extent that the state requirement imposes a 
higher level of performance and applies only to vehicles procured for 
the State's use. 49 U.S.C. 21461 sets forth a procedure for judicial 
review of final rules establishing, amending or revoking Federal motor 
vehicle safety standards. That section does not require submission of a 
petition for reconsideration or other administrative proceedings before 
parties may file suit in court.

g. Paperwork Reduction Act

    Under the Paperwork Reduction Act of 1995, a person is not required 
to respond to a collection of information by a Federal agency unless 
the collection displays a valid OMB control number. This final rule 
does not contain any collection of information requirements requiring 
review under the Paperwork Reduction Act.

h. National Technology Transfer and Advancement Act

    Section 12(d) of the National Technology Transfer and Advancement 
Act of 1995 (NTTAA) directs us to use voluntary consensus standards in 
our regulatory activities unless doing so would be inconsistent with 
applicable law or otherwise impractical. Voluntary consensus standards 
are technical standards (e.g., materials specifications, test methods, 
sampling procedures, and business practices) that are developed or 
adopted by voluntary consensus standards bodies, such as the 
International Organization for Standardization (ISO). The NTTAA directs 
us to provide Congress, through OMB, explanations when we decide not to 
use available and applicable voluntary consensus standards. The agency 
searched for, but did not find, voluntary consensus standards for use 
at this time.

i. Viewing Docket Submissions

    You may read the comments received by Docket Management at Room PL-
401, 400 Seventh Street, SW., Washington, DC, 20590 (telephone 202-366-
9324). You may visit the Docket from 10 a.m. to 5 p.m., Monday through 
Friday.

[[Page 37654]]

    You may also see the comments on the Internet. To read the comments 
on the Internet, take the following steps:
    (1) Go to the Docket Management System (DMS) Web page of the 
Department of Transportation (http://dms.dot.gov/).
    (2) On that page, click on ``search.''
    (3) On the next page (http://dms.dot.gov/search/), type in the 
four-digit docket number shown at the beginning of this document. 
Example: If the docket number were ``NHTSA-2002-1234,'' you would type 
``1234.'' After typing the docket number, click on ``search.''
    (4) On the next page, which contains docket summary information for 
the docket you selected, click on the desired comments. You may 
download the comments. However, since the comments are imaged 
documents, instead of word processing documents, the downloaded 
comments are not word searchable.
    Anyone is able to search the electronic form of all comments 
received into any of our dockets by the name of the individual 
submitting the comment (or signing the comment, if submitted on behalf 
of an association, business, labor union, etc.). You may review DOT's 
complete Privacy Act Statement in the Federal Register published on 
April 11, 2000 (Volume 65, Number 70; Pages 19477-78) or you may visit 
http://dms.dot.gov.

List of Subjects in 49 CFR Part 571

    Motor vehicle safety, Reporting and recordkeeping requirements, 
Tires, Incorporation by Reference.

PART 571--[AMENDED]

0
In consideration of the foregoing, NHTSA amends 49 CFR part 571 as set 
forth below.
0
1. The authority citation for Part 571 continues to read as follows:

    Authority: 49 U.S.C. 322, 30111, 30115, 30117 and 30166; 
delegation of authority at 49 CFR 1.50.


0
2. Section 571.5 is amended by renumbering the current paragraph 
(b)(10) as (b)(11) and adding a new paragraph (b)(10), to read as 
follows:


Sec.  571.5  Matter incorporated by reference

* * * * *
    (10) Child Restraint Systems Seat Assembly Drawing Package. Copies 
may be obtained by contacting: Leet-Melbrook, 18810 Woodfield Road, 
Gaithersburg, MD, 20879, telephone (301) 670-0090.
* * * * *

0
3. Section 571.213 is amended by:
0
a. Revising the definition of ``child restraint system'' in S4;
0
b. Adding S5(d);
0
c. Revising the introductory text of S5.1.2;
0
d. Adding S5.1.2.1 and S5.1.2.2;
0
e. Revising the table to S5.2.1.1(a);
0
f. Revising the introductory text of S5.2.1.2,
0
g. Revising S5.2.3.1 and S5.9(a);
0
h. Revising S6.1.1(a)(1), S6.1.1(b)(1) and S6.1.1(d), and the 
introductory text of S6.2.3;
0
i. Revising S7, and S9.1(c);
0
j. Adding S9.1(d), S9.1(e) and S9.1(f);
0
k. Revising S9.3, S10.2.1(b)(2), S10.2.1(c)(1)(i) introductory text, 
and S10.2.1(c)(2);
0
l. Revising Figure 1A; and,
0
m. Adding Figure 2A.
    The revised and added text and figures read as follows:


Sec.  571.213  Standard No. 213, Child restraint systems.

* * * * *
    S4. Definitions.
    Child restraint system means any device, except Type I or Type II 
seat belts, designed for use in a motor vehicle or aircraft to 
restrain, seat, or position children who weigh 30 kilograms (kg) or 
less.
* * * * *
    S5. Requirements.
* * * * *
    (d) Each child restraint tested with a part 572 subpart N dummy 
that is weighted to weigh 28.2 kg need not meet S5.1.2 and S5.1.3.
* * * * *
    S5.1.2 Injury criteria. When tested in accordance with S6.1 and 
with the test dummies specified in S7, each child restraint system 
manufactured before August 1, 2005, that, in accordance with S5.5.2, is 
recommended for use by children whose mass is more than 10 kg shall--
* * * * *
    S5.1.2.1 When tested in accordance with S6.1 and with the test 
dummies specified in S7, each child restraint system manufactured on or 
after August 1, 2005 shall'
    (a) Limit the resultant acceleration at the location of the 
accelerometer mounted in the test dummy head such that, for any two 
points in time, t1 and t2, during the event which are separated by not 
more than a 36 millisecond time interval and where t1 is less than t2, 
the maximum calculated head injury criterion (HIC36) shall not exceed 
1,000, determined using the resultant head acceleration at the center 
of gravity of the dummy head, ar, expressed as a multiple of g (the 
acceleration of gravity), calculated using the expression:
[GRAPHIC][TIFF OMITTED]TR18JN03.018

    (b) The resultant acceleration calculated from the output of the 
thoracic instrumentation shall not exceed 60 g's, except for intervals 
whose cumulative duration is not more than 3 milliseconds.
    S5.1.2.2 At the manufacturer's option (with said option irrevocably 
selected prior to, or at the time of, certification of the restraint), 
child restraint systems manufactured before August 1, 2005 may be 
tested to the requirements of S5 while using the test dummies specified 
in S7.1.2 of this standard according to the criteria for selecting test 
dummies specified in that paragraph. That paragraph specifies the 
dummies used to test child restraint systems manufactured on or after 
August 1, 2005. If a manufacturer selects the dummies specified in 
S7.1.2 to test its product, the injury criteria specified by S5.1.2.1 
of this standard must be met. Child restraints manufactured on or after 
August 1, 2005 must be tested using the test dummies specified in 
S7.1.2.
* * * * *
    S5.2 Force distribution.
* * * * *
    S5.2.1.1 * * *
    (a) * * *

                          Table to S5.2.1.1(a)
------------------------------------------------------------------------
                                                                 Height
                          Weight \1\                            \2\(mm)
------------------------------------------------------------------------
Not more than 18 kg..........................................        500
More than 18 kg..............................................       560
------------------------------------------------------------------------
\1\ When a child restraint system is recommended under S5.5 for use by
  children of the above weights.
\2\ The height of the portion of the system seat back providing head
  restraint shall not be less than the above.

* * * * *
    S5.2.1.2 The applicability of the requirements of S5.2.1.1 to a 
front-facing child restraint, and the conformance of any child 
restraint other than a car bed to those requirements, is determined 
using the largest of the test dummies specified in S7 for use in 
testing that restraint, provided that the 6-year-old dummy described in 
subpart I or subpart N of part 572 of this title is not used to 
determine the applicability of or compliance with S5.2.1.1. A front-
facing child restraint system is not required to comply with S5.2.1.1 
if the target point on either side

[[Page 37655]]

of the dummy's head is below a horizontal plane tangent to the top of--
* * * * *
    S5.2.3.1 Each child restraint system other than a child harness, 
manufactured before August 1, 2005, that is recommended under S5.5.2 
for a child whose mass is less than 10 kg and that is not tested with 
the Part 572 Subpart R dummy, shall comply with S5.2.3.
    S5.9 Attachment to child restraint anchorage system.
    (a) Each add-on child restraint anchorage system manufactured on or 
after September 1, 2002, other than a car bed, harness and belt-
positioning seat, shall have components permanently attached to the 
system that enable the restraint to be securely fastened to the lower 
anchorages of the child restraint anchorage system specified in 
Standard No. 225 (Sec.  571.213) and depicted in Drawing Package SAS-
100-1000 with Addendum A: Seat Base Weldment (consisting of drawings 
and a bill of materials), dated October 23, 1998, or in Drawing 
Package, ``NHTSA Standard Seat Assembly; FMVSS No. 213, No. NHTSA-213-
2003,'' (consisting of drawings and a bill of materials) dated June 3, 
2003 (incorporated by reference; see Sec.  571.5). The components must 
be attached by use of a tool, such as a screwdriver. In the case of 
rear-facing child restraints with detachable bases, only the base is 
required to have the components.
* * * * *
    S6.1.1 Test conditions.
    (a) Test devices.
    (1) Add-on child restraints.
    (ii) The test device for add-on restraint systems manufactured 
before August 1, 2005 is a standard seat assembly consisting of a 
simulated vehicle bench seat, with three seating positions, which is 
described in Drawing Package SAS-100-1000 with Addendum A: Seat Base 
Weldment (consisting of drawings and a bill of materials), dated 
October 23, 1998 (incorporated by reference in Sec.  571.5). The 
assembly is mounted on a dynamic test platform so that the center SORL 
of the seat is parallel to the direction of the test platform travel 
and so that movement between the base of the assembly and the platform 
is prevented.
    (ii) The test device for add-on restraint systems manufactured on 
or after August 1, 2005 is a standard seat assembly consisting of a 
simulated vehicle bench seat, with three seating positions, which is 
depicted in Drawing Package, ``NHTSA Standard Seat Assembly; FMVSS No. 
213, No. NHTSA-213-2003,'' (consisting of drawings and a bill of 
materials) dated June 3, 2003 (incorporated by reference; see Sec.  
571.5). The assembly is mounted on a dynamic test platform so that the 
center SORL of the seat is parallel to the direction of the test 
platform travel and so that movement between the base of the assembly 
and the platform is prevented.
* * * * *
    (b) * * *
    (1) Test Configuration I, are at a velocity change of 48 km/h with 
the acceleration of the test platform entirely within the curve shown 
in Figure 2 (for child restraints manufactured before August 1, 2005) 
or in Figure 2A (for child restraints manufactured on or after August 
1, 2005), or for the specific vehicle test with the deceleration 
produced in a 48 km/h frontal barrier crash.
* * * * *
    (d)(1) When using the test dummies specified in 49 CFR Part 572, 
subparts C, I, J, or K, performance tests under S6.1 are conducted at 
any ambient temperature from 19[deg] C to 26[deg] C and at any relative 
humidity from 10 percent to 70 percent.
    (2) When using the test dummies specified in 49 CFR Part 572, 
subparts N, P or R, performance tests under S6.1 are conducted at any 
ambient temperature from 20.6[deg] C to 22.2[deg] C and at any relative 
humidity from 10 percent to 70 percent.
* * * * *
    S6.2.3 Pull the sling tied to the dummy restrained in the child 
restraint system and apply the following force: 50 N for a system 
tested with a newborn dummy; 90 N for a system tested with a 9-month-
old dummy; 90 N for a system tested with a 12-month-old dummy; 200 N 
for a system tested with a 3-year-old dummy; or 270 N for a system 
tested with a 6-year-old dummy; or 350 N for a system tested with a 
weighted 6-year-old dummy. The force is applied in the manner 
illustrated in Figure 4 and as follows:
* * * * *
    S7 Test dummies. (Subparts referenced in this section are of part 
572 of this chapter.) S7.1 Dummy selection. Select any dummy specified 
in S7.1.1, S7.1.2 or S7.1.3, as appropriate, for testing systems for 
use by children of the height and mass for which the system is 
recommended in accordance with S5.5. A child restraint that meets the 
criteria in two or more of the following paragraphs in S7 may be tested 
with any of the test dummies specified in those paragraphs.
    S7.1.1 Child restraints that are manufactured before August 1, 
2005, are subject to the following provisions:
    (a) A child restraint that is recommended by its manufacturer in 
accordance with S5.5 for use either by children in a specified mass 
range that includes any children having a mass of not greater than 5 
kg, or by children in a specified height range that includes any 
children whose height is not greater than 650 mm, is tested with a 
newborn test dummy conforming to part 572 subpart K.
    (b) A child restraint that is recommended by its manufacturer in 
accordance with S5.5 for use either by children in a specified mass 
range that includes any children having a mass greater than 5 but not 
greater than 10 kg, or by children in a specified height range that 
includes any children whose height is greater than 650 mm but not 
greater than 850 mm, is tested with a newborn test dummy conforming to 
part 572 subpart K, and a 9-month-old test dummy conforming to part 572 
subpart J.
    (c) A child restraint that is recommended by its manufacturer in 
accordance with S5.5 for use either by children in a specified mass 
range that includes any children having a mass greater than 10 kg but 
not greater than 18 kg, or by children in a specified height range that 
includes any children whose height is greater than 850 mm but not 
greater than 1100 mm, is tested with a 9-month-old test dummy 
conforming to part 572 subpart J, and a 3-year-old test dummy 
conforming to part 572 subpart C and S7.2, provided, however, that the 
9-month-old dummy is not used to test a booster seat.
    (d) A child restraint that is recommended by its manufacturer in 
accordance with S5.5 for use either by children in a specified mass 
range that includes any children having a mass greater than 18 kg, or 
by children in a specified height range that includes any children 
whose height is greater than 1100 mm, is tested with a 6-year-old child 
dummy conforming to part 572 subpart I.
    (e) A child restraint that is manufactured on or after August 1, 
2005, and that is recommended by its manufacturer in accordance with 
S5.5 for use either by children in a specified mass range that includes 
any children having a mass greater than 22.7 kg, or by children in a 
specified height range that includes any children whose height is 
greater than 1100 mm, is tested with a part 572 subpart N dummy that is 
weighted to weigh 28.2 kg.
    S7.1.2 Child restraints that are manufactured on or after August 1,

[[Page 37656]]

2005, are subject to the following provisions.
    (a) A child restraint that is recommended by its manufacturer in 
accordance with S5.5 for use either by children in a specified mass 
range that includes any children having a mass of not greater than 5 
kg, or by children in a specified height range that includes any 
children whose height is not greater than 650 mm, is tested with a 
newborn test dummy conforming to part 572 subpart K.
    (b) A child restraint that is recommended by its manufacturer in 
accordance with S5.5 for use either by children in a specified mass 
range that includes any children having a mass greater than 5 but not 
greater than 10 kg, or by children in a specified height range that 
includes any children whose height is greater than 650 mm but not 
greater than 850 mm, is tested with a newborn test dummy conforming to 
part 572 subpart K, and a 12-month-old test dummy conforming to part 
572 subpart R.
    (c) A child restraint that is recommended by its manufacturer in 
accordance with S5.5 for use either by children in a specified mass 
range that includes any children having a mass greater than 10 kg but 
not greater than 18 kg, or by children in a specified height range that 
includes any children whose height is greater than 850 mm but not 
greater than 1100 mm, is tested with a 12-month-old test dummy 
conforming to part 572 subpart R, and a 3-year-old test dummy 
conforming to part 572 subpart P and S7.2, provided, however, that the 
12-month-old dummy is not used to test a booster seat.
    (d) A child restraint that is recommended by its manufacturer in 
accordance with S5.5 for use either by children in a specified mass 
range that includes any children having a mass greater than 18 kg, or 
by children in a specified height range that includes any children 
whose height is greater than 1100 mm, is tested with a 6-year-old child 
dummy conforming to part 572 subpart N.
    (e) A child restraint that is manufactured on or August 1, 2005, 
that is recommended by its manufacturer in accordance with S5.5 for use 
either by children in a specified mass range that includes any children 
having a mass greater than 22.7 kg or by children in a specified height 
range that includes any children whose height is greater than 1100 mm 
is tested with a part 572 subpart N dummy that is weighted to weigh 
28.2 kg.
    S7.1.3 Voluntary use of alternative dummies. At the manufacturer's 
option (with said option irrevocably selected prior to, or at the time 
of, certification of the restraint), child restraint systems 
manufactured before August 1, 2005 may be tested to the requirements of 
S5 while using the test dummies specified in S7.1.2 according to the 
criteria for selecting test dummies specified in that paragraph. Child 
restraints manufactured on or after August 1, 2005, must be tested 
using the test dummies specified in S7.1.2.
* * * * *
    S9.1 Type of clothing.
* * * * *
    (c) 12-month-old dummy (49 CFR Part 572, Subpart R). When used in 
testing under this standard, the dummy specified in 49 CFR part 572, 
subparts R, is clothed in a cotton-polyester based tight fitting sweat 
shirt with long sleeves and ankle long pants whose combined weight is 
not more than 0.25 kg.
    (d) Hybrid II three-year-old and Hybrid II six-year-old dummies (49 
CFR part 572, subparts C and I). When used in testing under this 
standard, the dummies specified in 49 CFR part 572, subparts C and I, 
are clothed in thermal knit, waffle-weave polyester and cotton 
underwear or equivalent, a size 4 long-sleeved shirt (3-year-old dummy) 
or a size 5 long-sleeved shirt (6-year-old dummy) having a mass of 
0.090 kg, a size 4 pair of long pants having a mass of 0.090 kg, and 
cut off just far enough above the knee to allow the knee target to be 
visible, and size 7M sneakers (3-year-old dummy) or size 12\1/2\M 
sneakers (6-year-old dummy) with rubber toe caps, uppers of dacron and 
cotton or nylon and a total mass of 0.453 kg.
    (e) Hybrid III 3-year-old dummy (49 CFR Part 572, Subpart P). When 
used in testing under this standard, the dummy specified in 49 CFR part 
572, subpart P, is clothed in thermal knit, waffle-weave polyester and 
cotton underwear or equivalent, a size 4 long-sleeved shirt (3-year-old 
dummy) or a size 5 long-sleeved shirt (6-year-old dummy) having a mass 
of 0.090 kg, a size 4 pair of long pants having a mass of 0.090 kg, and 
cut off just far enough above the knee to allow the knee target to be 
visible, and size 7M sneakers with rubber toe caps, uppers of dacron 
and cotton or nylon and a total mass of 0.453 kg.
    (f) Hybrid III 6-year-old dummy (49 CFR Part 572, Subpart N) and 
Hybrid III weighted 6-year-old dummy. When used in testing under this 
standard, the dummy specified in 49 CFR part 572, subpart N, weighted 
and unweighted, is clothed in a light-weight cotton stretch short-
sleeve shirt and above-the-knee pants, and size 12\1/2\ M sneakers with 
rubber toe caps, uppers of dacron and cotton or nylon and a total mass 
of 0.453 kg.
* * * * *
    S9.3 Preparing dummies. (Subparts referenced in this section are of 
Part 572 of this chapter.)
    S9.3.1 When using the test dummies conforming to Part 572 Subpart 
C, I, J, or K, prepare the dummies as specified in this paragraph. 
Before being used in testing under this standard, dummies must be 
conditioned at any ambient temperature from 19[deg] C to 25.5[deg] C 
and at any relative humidity from 10 percent to 70 percent, for at 
least 4 hours.
    S9.3.2 When using the test dummies conforming to Part 572 Subparts 
N (weighted and unweighted), P, or R, prepare the dummies as specified 
in this paragraph. Before being used in testing under this standard, 
dummies must be conditioned at any ambient temperature from 20.6[deg] C 
to 22.2[deg] C and at any relative humidity from 10 percent to 70 
percent, for at least 4 hours.
* * * * *
    S10.2.1 * * *
    (b) * * *
    (2) When testing rear-facing child restraint systems, place the 
newborn, 9-month-old or 12-month-old dummy in the child restraint 
system so that the back of the dummy torso contacts the back support 
surface of the system. For a child restraint system which is equipped 
with a fixed or movable surface described in S5.2.2.2 which is being 
tested under the conditions of test configuration II, do not attach any 
of the child restraint belts unless they are an integral part of the 
fixed or movable surface. For all other child restraint systems and for 
a child restraint system with a fixed or movable surface which is being 
tested under the conditions of test configuration I, attach all 
appropriate child restraint belts and tighten them as specified in 
S6.1.2. Attach all appropriate vehicle belts and tighten them as 
specified in S6.1.2. Position each movable surface in accordance with 
the instructions that the manufacturer provided under S5.6.1 or S5.6.2. 
If the dummy's head does not remain in the proper position, tape it 
against the front of the seat back surface of the system by means of a 
single thickness of 6 mm-wide paper masking tape placed across the 
center of the dummy's face.
    (c)(1)(i) When testing forward-facing child restraint systems, 
extend the arms of the 9-month-old or 12-month-old test dummy as far as 
possible in the upward vertical direction. Extend the legs of the 9-
month-old or 12-month-old test dummy as far as possible in the forward 
horizontal direction, with the dummy

[[Page 37657]]

feet perpendicular to the centerline of the lower legs. Using a flat 
square surface with an area of 2,580 square mm, apply a force of 178 N, 
perpendicular to:
* * * * *
    (2) When testing rear-facing child restraint systems, extend the 
dummy's arms vertically upwards and then rotate each arm downward 
toward the dummy's lower body until the arm contacts a surface of the 
child restraint system or the standard seat assembly in the case of an 
add-on child restraint system, or the specific vehicle shell or the 
specific vehicle, in the case of a built-in child restraint system. 
Ensure that no arm is restrained from movement in other than the 
downward direction, by any part of the system or the belts used to 
anchor the system to the standard seat assembly, the specific shell, or 
the specific vehicle.
* * * * *
BILLING CODE 4910-59-P
[GRAPHIC][TIFF OMITTED]TR18JN03.012


[[Page 37658]]


[GRAPHIC][TIFF OMITTED]TR18JN03.013


    Issued on: June 4, 2003.
Jeffrey W. Runge,
Administrator.
[FR Doc. 03-14425 Filed 6-23-03; 8:45 am]
BILLING CODE 4910-59-C