EPA-AA-IMS/81-13
                          Recommendations  Regarding  the
                      Selection of Idle Emission Cutpoints
                     for  Inspection  and Maintenance  Programs
               Requiring Only Carbon Monoxide Emission Reductions
                                       by
                                 Susan Vintilla
                          Inspection/Maintenance Staff
                                    May,  1981
                                     Notice
This Report  does not  necessarily  represent  final EPA decisions  or  positions.
It is intended to present  technical analysis  of  the  issue  using  data which are
currently  available.   The  purpose  in  the  release  of   such  reports  is  to
facilitate the  exchange of technical information  and  to  inform  the  public  of
technical  developments which  may  form  the  basis  for a  final   EPA  decision,
position or regulatory  action.
                         Inspection and Maintenance
                      Emission Control Technology Division
                  Office of Mobile Source Air Pollution Control
                       Office of Air, Noise, and Radiation
                      U.S. Environmental Protection Agency

-------
Preface

This report  is  a  supplement to EPA-AA-IMS/81-1,  "Recommendations Regarding the
Selection of  Idle  Emission Inspection Cutpoints  for Inspection and Maintenance
Programs"[1]*,  which  was  developed as a part  of  EPA's  Model Program guidance.
The original  report  described methods for selecting HC and  CO cutpoints given
an  idle  emission data  base,  and recommended  specific  HC  and CO  cutpoints  by
model  year  groupings  for desired  program  failure  rates.   The  recommended
cutpoints  in the  original report  were intended  for   use  in I/M  programs  in
areas which  require only  HC,  or both HC and CO,  emission  reductions  in order
to  attain  the  National  Ambient Air Quality  Standard  for  ozone, or  ozone and
CO.   This  report  Is  intended  to  aid I/M  programs  which  only  require  CO
reductions  in  the  selection  of cutpoints,   prediction of  failure  rates,  and
estimation   of   I/M  emission  benefits.   Its  recommendations  are  therefore
applicable  to  I/M  in areas that .have -already  attained the  ozone  standard  or
that are projected  to attain  the ozone  standard by 1982 without  I/M.
 ^Numbers  in brackets refer to a list  of  references at  the end of the report.

-------
1.0 INTRODUCTION

1.1 Previous EPA Outpoint Selection Guidance

The  recommendations  presented  in  Reference  1   for  pre-1981  vehicles  were
derived  froa a data  base  of vehicles which  had  undergone testing of  idle  HC
and  CO emissions.   Failure rates  for given  cutpoints  were  predicted  for  5
different  technology  types  corresponding  to model  year  groups  (pre-1968,
1968-1971,  1972-1974,  1975-1979 and 1980) by  a computer  model which  accounted
for  the  increase  in  idle  HC  and  CO  scores  with mileage.  Failure  rates  for
1981  and  future  vehicles  were  predicted   differently,  due  to  their  more
sophisticated   emission control  technology.   A  high   proportion  of  these
vehicles  will  have  computerized  control of  the  air-fuel  mixture,  which  is
expected to  make  them less  sensitive to cutpoint  variations.

Failure  rates  for  each  model year  grouping  in  the  previous  report  were
presented as the  total  percent of  tested  vehicles which  fail the idle test for
HC  and/or  CO,  given  certain  sets  of  HC  and  CO  cutpoints.    This  report will
recommend CO cutpoints for desired failure rates  for  CO  idle  tests only.  (As
explained  below,  EPA has   determined  that HC cutpoints   are  not  crucial  for
deriving a  large CO-only emission benefit.)   Naturally,  some of these vehicles
would  also   fail  HC  idle  tests but it  is not the  purpose  of this  report  to
detail  those  failures.  Failure  rate predictions  in this  report will only
apply  to  the  first   functioning  year  of an I/M  program;  as  explained  in
Reference  1, failure  rates  for the following years  can  be estimated  by -data
collected from the actual  operation of the program.

1.2  Major Points  Discussed  In  This  Report

The  following questions will be answered:

     1) What, if any,  should be the cutpoint  locus; i.e.,  should there be an HC
     cutpoint in addition  to a  CO  cutpoint?   If  so,  what multiple of  the  CO
     cutpoint should  it be?

     2) How  can states predict  failure  rates  from  specific  CO cutpoints?

     3) How  can the emission benefits be  calculated?

-------
2.0 SHOULD AN HC OUTPOINT BE USED?

2.1 Pre-1981 Vehicles

EPA's investigation of CO benefits  from  HC-only  failures  of pre-1981 vehicles
has  shown  that they  typically amount  to a  small  percent of  total  idle test
failures,  and  that  they  receive   essentially  no  CO  emission  benefit  from
repair.  The  lack  of a CO benefit is due  to  the  fact  that these vehicles' FTP
CO emission levels  at  inspection  are already low (as their ability to pass the
I/M  CO  outpoint would suggest) and  the fact that  no  CO-oriented repairs need
to be  performed for  these  vehicles  to pass  reinspection.  Table  1  shows the
as-received FTP CO emissions and  the change  in  CO emissions for vehicles from
repair of HC-only  failures  in  EPA's  Portland  Study  Sample.


                                      Table 1

                         FTP CO Emission Change Resulting
                    From Maintenance  On HC-Only Failures From
  -v                     Elements II and III and Test  Group 7
                            of  the Portland I/M Fleet


                                        As-received
               Percent of idle  test       FTP CO         Percent  change  in FTP CO
  Vehicle       failures  which  are      emissions       emissions  from HC-only
model year      HC-only  failures          (g/mi)          failures  after repair

1972-1974             6.3                27.98                   - 2.2
1975-1977             7.4                16.70                   +22.2


As can  be seen, there was actually a small increase (3.7  g/mi)  in  CO  emissions
after  repair for the 1975-1977 model ..year group.

2.2  1981 and Future Vehicles

Emission control  system  failures of the  1981 and  future  fleet can be divided
into two  categories:  1) failures of the closed-loop  computer control system,
which  result in rich  operation and  high  HC  and  CO emissions 2)  other failures
which  are also common to older vehicles, caused  by tampering,  malmaintenance,
misfueling, and ignition-related problems.   As explained  in Reference 1,  it is
 the  failures of the  first  category  which are expected  to  be the main cause of
 1981  and  later  vehicles   emitting significantly  above  the  standard.   The
 emission factors  for  this  failure category,  as  predicted in MOBILE2  and. shown
 in  Table  2,  demonstrate that  both  HC and CO emissions  will  be  exceptionally
 high [4].  A CO-only  cutpoint  is  expected to identify  essentially  all of  these
 vehicles  [3].   Failures of  the  second  category  will  be similar  to  those of
 pre-1981 vehicles, and  so  the same  arguments for  the suitability  of  a CO-only
 cutpoint are applicable.

-------
                                        5

                                     Table 2
  . '             Emission Rates for 1981 and Future Vehicles  Which
              Experience Failure of the Closed-Loop Control System

                                           New Vehicle
                 Pollutant             Emission Rate (g/mi)

                    HC                         3.74
                    CO                       107.36

2.3 Undesirable Consequences  of Emphasizing HC Idle Failures

EPA  has  found  that   there   are  certain  characteristics   associated  with  HC
failures  on the idle  test  which areas that need only CO  reductions  from the
I/M program would  logically wish to  avoid  if  possible.  Cutpoint sets  for HC
and  CO which  emphasize HC,  for  example,  are accompanied  by higher  rates  of
errors  of  commission  (vehicles  which  fail  the  idle   test  but actually  meet
federal  emission  standards).   Also,   idle  failures for  HC  often  necessitate
cftstly  ignition repairs, which  could needlessly  drive up the average  repair
cost  in  the  I/M  program  with little  resulting  CO  emission  reduction.   Yet
another  issue related  to  the choice  of idle  inspection  type is  that of  fuel
economy benefit from repair of failed vehicles.   EPA has  found that  there  is a
net fuel  economy benefit of  0.8%  for pre-1981 vehicles repaired by mechanics
who have  been educated in conventional training  programs.   However,  this  fuel
economy benefit rises  to 4%  for  these same vehicles when  mechanics  only  adjust
carburetors to a specific idle CO  level  [2].  This would  indicate that  an I/M
program where repairs  for pre-1981 vehicles are  limited  to carburetor adjust-
ments  would yield  the  maximum CO and  fuel  economy benefit,  at low  cost.   By
using  only  a CO  outpoint,  an I/M  program can  effectively  limit   repairs  to
carburetor  adjustments [3].   (The  fuel  economy  benefit   for  1981  and  later
vehicles  is not expected to be influenced by mechanic training [2].)

In summary, a program  which consists  of a  CO-only idle inspection will deliver
full  CO emission reduction benefits,  and  may  actually deliver a greater  fuel
economy   benefit   for   pre-1981  vehicles  since   repairs  can  be  limited  to
carburetor  adjustments.  Also,  a CO-only  program  operating  at  a  particular
failure  rate will  deliver  a greater CO benefit than an HC/CO program operating
at Chat  same failure rate  [3].

Additional  discussion  of the  ramifications of using only a CO  cutpoint  can be
found  in  Reference 3.

-------
3.0 HOW CAN STATES PREDICT  FAILURE RATES FROM  SPECIFIC CO OUTPOINTS?

As discussed in Reference  1,  EPA recommends  using data from a pilot program or
a  mandatory inspection/voluntary  testing program  in the  intended geographic
location,  for  the most  accurate estimation of  failure  rates.   For situations
where this is not possible, EPA developed  a model to predict failure rates for
various cutpoints,  based on  data collected  from EPA's I/M Demonstration Vans,
EPA's Emission  Factor Program and New Jersey's  I/M  program  [1],   Tables 3 and
4  present  predicted  minimum,  maximum  and  mean  failure  rates   for  CO-only
cutpoints,  based on  this  model.  Table  3  is 'for  programs  beginning  1/1/82;
Table 4  for  those starting a year later.  The minimum is  the failure  rate for
the  most  recent  and  therefore the  youngest  and cleanest  model  year  in the
group;  the maximum  is  the  failure  rate  for  the  oldest  and therefore  the
dirtiest  model  year;  and  the mean is the  average  for all  the model  years in
the  group,  taking the national average, registraton distribution of model years
within model year groups at the projected  time into  account.

Based on  these  tables, EPA has estimated the CO cutpoints that would be needed
at  the  beginning of  an I/M  program  for  a variety  of failure  rates.   Table' 5
presents  these  cutpoints   for  I/M  programs beginning  1/1/82 and  1/1/83.   As
dxscussed in Reference  1, failure rates  for pre-1981  vehicles  should remain
approximately   constant  during  the   first  year  of  a  program  if  vehicle
inspections  fall on  vehicle  birthdays.   However,  EPA's  model predicts   that
vehicle model year group failure rates at the second 'inspection will decrease,
for  example, 4-14%  for a given set of cutpoints designed to achieve an initial
failure  rate of  35%  [1].   Also,  the data  base used  for  prediction  of   1980
model  year  failure   rates  was  limited,  and  so  these   predictions  may  be
inaccurate [1].  Therefore,  states will probably need to adjust  cutpoints for
pre-1981  vehicles yearly   if  they wish  to  maintain a  constant  failure rate.
Also, it  may be necessary  for a state to revise cutpoints within the first few
months  of the  start  of a   program,  to correct for  differences  caused  by local
influences.              •        ' '

Failure  rates   for 1981  and later vehicles will  be  less sensitive to cutpoint
variations,  due  to   the  nature of  "the  typical failure  of  their  computer-
controlled emission   control  systems.   As  mentioned above,  data  from these
vehicles   indicate  that  these  failures  will  cause  a  gross  increase in CO
emissions, which will be easily identified by  any reasonable CO cutpoint.

-------
                                                           Table 3

                                            Initial Model Year  Group  Failure Rates
                                              Predicted by  EPA  CO  Outpoint Model
                                               At  I/M  Program Start Date  1/1/82
                                             Model Year Group Failure Rates  (%)
   CO  Idle
 CutpoinC  (%)
     Pre-1968
Mini/  Max2/  Mean3/
    1968-71
Min   Max   Mean
    1972-74
Min   Max   Mean
    1975-79
Min   Max   Mean
1980 4/ 1981+ 5/
Mean
1.0
1.2
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
10.0
Registration
Fraction


64.3
60.9
55.3
49.6
43.9
39.1
36.3
30.7
25.0
20.5
15.8
11.1
1.8



64.6
61.5
55.3
51.8
45.1
39.8
36.9
31.8
26.2
22.2
17.8
13.5
2.5
.038


64.4
61.2
55.3
.50.7
44.4
39.6
36.6
31.4
25.7
21.4
16.8
12.4
2.2



58.4
53.4
48.9
43.0
37.2
32.7
27.2
22.1
19.6
14.9
11.0





58.4
54.1
50.0
44.4
37.8
33.9
28.6
24.3
20.4
15.9
13.3


.110


58.4
53.9
49.0
43.9
37.8
33.0
27.8
23.2
19.7
15.3
11.9



74.4
69.2
62.6
55.7
51.0
45.4
40.8
36.2
31.6
26.3
22.1
18.6
15.5
11.9
9.5



77.4
71.6
65.1
58.5
53.6
48.3
43.6
47.7
35.0
29.2
25.0
21.0
17.4
14.7
11.5


.184
75.7
70.5
63.8
57.0
51.9
46.8
42.0
47.3
32.8
27.8
23.6
19.7
16.2
13.1
10.4



46.3
39.6
34.9
30.8
27.2
. 23.7
20.6
17.3
14.5
12.6
10.4
8.7
7.0
5.7
4.1



65.2
47.5
42.3
37.9
33.5
30.2
26.7
23.7
20.9
18.3
15.4
13.3
11.6
9.9
8.4


.449
53.5
44.0
38.7
34.5
30.6
27.2
23.9
20.9
18.0
15.4
13.0
11.2
9.5
7.9
6.4



33.8
1.0-3.0
30.0
25.6
--21. 7
18.7
14.9
12.7
10.5
8.6
6.6
5.2
3.7
2.7
1.9
1.0


.095 .124
(1) Minimum model year failure rate within model year groups.
(2) Maximum model year failure rate within model year groups.
(3) Average failure rate in model year groups weighted by projected model year registration fractions.
(4) EPA predictions for 1980 vehicles are baaed on limited data and therefore may be less accurate
    than those for pre-1980 vehicles.
(5) EPA recommends use of a 207(b) short test standard of 1.2% CO.
    EPA expects that the failure rate from these cutpoints will never exceed 1-3%.

-------
                                                           Table 4

                                            Initial  Model  Year  Group  Failyre  Rates
                                              Predicted  by EPA  CO Cutpoin't Model
                                               At  I/M Program Start Date  1/1/83
                                             Model Year Group Failure Rates  (%)
 Outpoint  (%)
   CO  Idle
     Pre-1968
Mini/  Max2/  Mean3/
    1968-71
Min   Max   Mean
    1972-74
Min   Max   Mean
    1975-79
Min   Max   Mean
1980 47 1981+ 5_/
Mean
1.0
1.2
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5 •'
6.0
6.5
7.0
7.5
8.0
8.5
10.0
Registration
Fraction


64.3
60.9
55.3
49.6
43.9
39.6
36.3
31.4
25.4
21.1
16.2
11.7
2.1



64.6
61.5
55.3
51.8
45.1
39.8
36.9
31.8
26.2
22.2
17.8
13.5
2.5
.027


64.5
61.3
55.3
51.2
44.7
39.7
36.7
31.7
26.0
21.8
17.2
12.4
2.4



58.4
54.1
48.9
44.4
37.8
32.7
27.8
23.1
19.6
15.3
11.9





58.4
54.1
50.0
45.1
38.7
33.9
28.6
24.5
20.4
17.0
13.3


.084


58.4
54.1
49.2
44.5
38.2
33.2
28.1
23.8
19.9
15.7
12.4



75.7
71.1
64.2
57.2
51.4
47.3
41.9
38.1
32.4
28.2
24.1
19.9
16.2
13.2
10.4



79.1
72.9
67.0
60.2
54.9
49.5
44.5
40.1
36.1
30.0
25.7
21.5
18.6
15.1
11.9


.152
77.1
71.7
65.2
58.5
53.0
48.2
43.1
38.9
34.2
29.0
24.8
20.7
17.2
14.2
11.1



48.1
42.3
37.2
33.3
29.1
25.6
• 22.2
19.1
16.5
13.9
11.9
10.1
8.4
6.6
5.5



67.1
50.6
44.1
38.6
34.5
31.7
28.0
25.0
21.7
19.1
16.5.
14.2
12.6
10.5
9.2


.447
56.9
45.8
40.2
35.7
31.8
28.5
25.1
22.1
19.1
16.5
14.0
12.1
10.4
8.7
7.2



35.8
1.0-3.0
32.0
27.4
24.6
20.9
18.2
14.5
12.6
10.5
8.7
7.1
5.7
4.2
3.0
2.2


.073 .217
(1) Minimum model year.failure rate within model year groups.
(2) Maximum model year failure rate within model year groups.
(3) Average failure rate in model year groups weighted by projected model year registration fractions,
(4) EPA predictions for  1980 vehicles are based on limited data and therefore may be less accurate
    than those for pre-1980 vehicles.
(5) EPA recommends use of a 207(b) short test standard of 1.2% CO.
    EPA expects that the failure rate from these cutpoints will never exceed 1-3%.

-------
                                                           Table 5

                                              EPA Recommended  I/M  CO  Outpoints

                 Outpoints predicted to give constant failure rates among pre-1981 vehicles for I/M starting
                           1/1/82 and 1/1/83 in initial year inspections by EPA CO cutpoint model.
I/M
Start
Date
1/1/82




1/1/83




Nominal
Failure
Rate (%)

20
25
30
35

20
25
30
35

Pre-1968
CO (%)

7.5
7.0
6.5
6.0

7.5
7.0
6.5
6.0
Recommended
1968-71
CO (Z)

7.0
6.5
6.0
5.5

7.0
6.5
6.0
5.5
CO Outpoints
1972-74
CO (%)

6.5
6.0
5.5
5.0

6.5
6.0
5.5
5.0
for Model
1975-79
CO (%)

4.5
4.0
3.0
2.5

5.0
4.0
3.5
2.5
Year Groups
1980 I/
CO (%)

3.0
2.0
1.5
1.0

3.0
2.5
1.5
1.0

198H- 2/
CO (%)

1.2
- 1.2
1.2
1.2

1.2
1.2
1.2
1.2
If EPA1s predictions for 1980 vehicles are based on limited data and therefore may be less accurate
   than those for pre-1980 vehicles.                .
21 All I/M programs should use the 207(b) cutpoint for 1981 and later vehicles for all program years,
   that the failure rate from this cutpoint will never exceed 1-3%.
EPA expects

-------
                                        10

4.0 HOW CAN THE  EMISSION  BENEFITS BE CALCULATED?

The MOBILE2  emission  factors  model  can be  used,  with modifications.   Orice a
state has  decided on the failure rate  desired,  it  must choose the appropriate
CO outpoints  for each pre-1981 model  year group based on data  from  the local
vehicle  sample,  or based  on  the recommendations  in  this  report.   These  CO
cutpoints  should then be matched to appropriate hypothetical HC cutpoints from
EPA's  locus  pairs  as described  in  Table   1  of  Reference  1.    These  HC/CO
cutpoint pairs  should then  be  applied  to the local  data sample to find a total
failure  rate  for pre-1981 vehicles;  or they should be  used  to look up a total
failure  rate among pre-1981  vehicles  in Tables  3 or 4 of  Reference  1,  by
weighting  the   failure   rate   for  each  pre-1981  model  year  group  by  its
registration  fraction.   This  failure  rate should  then be used in MOBILE2  as
the stringency  for the I/M program for pre-1981  vehicles.   For 1981  and later
vehicles,  the   effect  of   an  I/M  program is  calculated   by  inputting  an
identification  (ID) rate instead of a  stringency rate.  The ID rate describes
the  fraction of  gross  HC and  CO  emitters  (caused  by  closed-loop  system
failure,  as  described in  Reference  4) which  will be identified by  the  I/M
test.   If  the  basic idle-in-neutral  test  will  be  used  for  an area's  I/M
program,  then a 50% ID rate should be  inputted into MOBILE2.  (This ID rate is
th"e  default  value  when   the input  flag IMFLG=1, therefore, only  the pre-1981
stringency value needs  to  be  entered  in this case.)  .If  the  I/M  program will
use a  two—speed  idle or  loaded mode test  and will  pass and  fail on both modes
of  the  test, then  an ID  rate  of  70% should  be  entered  for  1981  and later
vehicles.

Presented below  is a  sample calculation for a  hypothetical  area which plans a
CO-only I/M  program and which  will  use  the idle-in-neutral   test.   The area
wants  a failure  rate  of  25% for  the first year  of  its I/M program, which will
begin  on January 1,  1983.   Based on Table  5 of  this  report,  the  following CO
cutpoints should  be  used  for  the  first  year  of the  program  for pre-1981
vehicles:


                    Vehicle model year     CO  idle cutpoint, %

                       pre-1968                 7.0
                       1968-1971                6.5
                       1972-1974                6.0
                       1975-1979                4.0
                       1980                     2.5


These  CO  cutpoints are  then  matched   to their  hypothetical -EC cutpoints from
Table  1 of Reference  1,  to yield the following pairs:


          Vehicle model year      HC (ppm)              C0(%)

              pre-1968              700                  7.0
              1968-1971             650                  6.5
              1972-1974             600                  6.0
              1975-1979             400       '•          4.0
           "  1980                 275                  2.5

-------
                                        11

The HC/CO  cutpoint pairs are  then  used to look up  the  failure  rates for each
model  year  group  in  Table 4 of  Reference  1.   The failure  rates  are  then
weighted by  the  registration fraction of each model year group.  Note that the
registration fractions  listed  in  Table 4 must be normalized to account for the
absence of the 1981  and later  vehicles.
                                     Mean             Registration
          Vehicle model  year       Failure rate            fraction

             pre-1968             .   52.0             .027/.783 = .035
             1968-1971                38.9             .084/.783 = .107
             1972-1974                34.5             .152/.783 = .194
             1975-1979                37.1             .447/.7S3 = .571
             1980                     35.2             .073/.783 = .093
The   failure   rates  are  then  multiplied   by   their   respective  registration
fractions  and summed to yield  the  total  theoretical failure rate for pre-1981
vehicles:
            Mean failure                  Registration
            	rate                        fraction

                52.0            x             .035
                38.9            x             .107
                34.5            x             .194
                37.1            x             .571
                35.2            x             .093
                                 Total failure rate                  37.12
 Therefore,  a  stringency  of 37  will  be  used  for  pre-1981  vehicles  in the
 MOBILE2 program,  and  the default  ID rate  of  50% will  be  used  for  1981 and
 later  vehicles.   Comparison of  the  CO composite  emission factor  from this
 MOBILE2 run with the CO  composite emission factor  from a "no-I/M"  run will
 enable  the  user  to calculate  the  CO  emission  benefit for  this particular
 program.  (The  HC emission  factors  from these MOBILE2  runs  are  meaningless,
 and should be discarded.)

-------
                                       12

5.0 CONCLUSIONS

CO  cutpoints  are  sufficient  to  obtain  essentially  all  of  the  normal  CO
emission reduction  benefit  possible  from an  I/M program,  and  a CO-only program
operating  at  a given  failure rate will deliver  a  greater  CO benefit  than  an
HC/CO  program operating  at that  same  failure rate.   Cutpoints  needed  for  a
desired initial program  failure rate can be  chosen from the recommendations  in
this  report,  or  can be  determined  from analysis  of the  results  of  a  local
trial  I/M  program.   Cutpoints for  future program years may need  to  be altered
to insure  a constant  failure  rate,  and  cutpoint selection should be  based  on
results  from  the  first  year  of  the program.   CO  emission  benefits can  be
calculated  from  MOBILE2,  but  not  in  the  same  manner  as  for  an  HC/CO  I/M
program.   The  chosen CO cutpoints   for  each  pre-1981  model year  group  are
matched  to  hypothetical  HC  cutpoints  from  the  recommended  locus  pairs  in
Reference  1;  these pairs  are  used to  calculate   a total  failure  rate  for
pre-1981  vehicles  which  is   input  as  the  program stringency  in  MOBILE2.
Effectiveness  of  the I/M program for 1981  and later vehicles  is  represented  in
the input  by  an ID  rate  of 50 or 70 percent, depending on the type of emission
test  used.   CO emission factors  from MOBILE2  are  used  to calculate  emission
benefits,  and  the corresponding HC emission  factor outputs are ignored.

-------
                                       13

References


[1] U.S. EPA.  January,  1981.   Recommendations  Regarding the Selection of Idle
    Emission  Inspection  Outpoints  for  Inspection  and Maintenance  Programs.
    EPA-AA-IMS/81-1.

[2] U.S. EPA.  April,  1981.   Update on the  Fuel Economy Benefits of Inspection
    and Maintenance Programs.  EPA-AA-IMS/81-10.

[3] U.S.  EPA.    April,   1981.    Low-Cost  Approaches   to  Vehicle  Emissions
    Inspection and Maintenace.  EPA-AA-IMS/81-7.

[4] U.S.  EPA.   November,  1980.   Derivation   of  1981  and  Later  Light  Duty
    Vehicle   Emission   Factors   for   Low   Altitude,   Non-California   Areas.
    EPA-AA-IMS/80-8.

-------