EPA-AA-SDSB-79-12
Technical Report
Prediction of U.S. Annual Fuel Consumption
by Passenger Automobiles
by
Tamara Ward
Glenn Thompson
NOTICE
Technical Reports do not necessarily represent final EPA decisions
or positions. They are intended to present technical analysis of
issues 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 devel-
opments which may form the basis for a final EPA decision, position
or regulatory action.
Standards Development and Support Branch
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
Office of Air, Noise and Radiation
U.S. Environmental Protection Agency
-------
I. Introduction
The annual fuel consumption of U.S. passenger vehicles is an
area of present and increasing concern. In order to assess EPA
programs which may affect national fuel consumption, it is impor-
tant to be able to predict future fuel economy trends and possible
modifications of these trends. .
This report presents a computer model which can be used to
predict trends in U.S. passenger vehicle fuel consumption. While
this model is relatively simple to allow easy use, it is suffi-
ciently detailed to provide accurate relative predictions of
different fuel conservation strategies. The model methodology was
developed primarily to investigate the fuel consumption implica-
tions of various applications of tire technology. During the
course of the model development it was decided to present this
material as a separate report to facilitate use of this material .in
other fuel consumption prediction efforts.
The. actual prediction model used in this is quite simple.
However, the mathematical methodlogy of the model can be easily
extended to address much more complex applications. In general,
the limiting condition in using the model will be the availability
of detailed input information.
II. Discussion
~ " The development of a useful model requires two tasks: first,
the model must be chosen, and second) there must be a literature
search to provide the necessary model input parameters. While the
first task may be conceptually more difficult, the second is often
more time consuming, and is essential for applications of the
model. Consequently, both the model and the development of cur-
rently suitable input parameters is discussed in the following
subsections. The final section presents the predicted annual fuel
consumption for 1975 through 1985, and compares the predicted
values for current years with reported data.
A. The Model
The fuel economy prediction model was chosen as:
TFCON. = i (VMIX..)(MIT..)(FC.) (1)
1 j ^J 1J J
where:
TFCON. = the total annual fuel consumed in the i
year.
-------
VMIX.. = the vehicle mix of j type vehicles in the year i.
MIT.. = the annual vehicle miles traveled by j type vehicles
in the year i.
FC. = the average fuel consumption of j type vehicles.
This model is basically the model presented by H.H. Gould and
A.C. Mallioris of DOT._1_/* While the model is simple, it has great
versatility, since the vehicle mix paramter j can represent . as
detailed subdivision of the vehicle population as is desired. For
example, a k subscript could be introduced to each of the para-
meters VMIX, MIT and FC to represent different tire types.
The fact that the model equation (1) is so simple, yet ex-
tremely powerful means that much of the information content of the
model resides in the values of the model paramaters VMIX, MIT, and
FC. The subsequent section discusses how the values to these para-
meters were chosen and presents the values used in this analysis.
B. Values of the Model Parameters
In this analysis only the vehicle model year was considered as
a subdivding parameter of the vehicle mixture parameter, VMIX.
This approach was chosen because of the difficulty in obtaining
more detailed information on the distribution of the vehicle
population.. With this choice of the vehicle subdivision parameter,
the quantities which must be obtained or constructed for each of
the i years of interest, are VMIXj. MIT and FC., that is .the
number of vehicles of model year j existent in the calendar year i,
the annual miles traveled by vehicles of model year j in calendar
year i, and the average fuel economy of vehicles of model year j.
In general, this desired information is available for the past
.ten years. It is believed that these data can be accurately
extrapolated for approximately the same time period into the
future. Therefore, it was decided to predict the total annual fuel
consumption from the present time until 1985. Consequently, the
required parameters must be known or estimated for model years 1978
to 1985, inclusive.
1. The Vehicle Population Distribution
The distribution of vehicles by model year in past years is
readily available, usually from data compiled by R.L. Polk from
state registration lists.2j Table 1 gives the currently available
model year distribution. The first sub-task then is to take this
distribution and use it to predict the vehicle population distribu-
tions for each year until 1985. This was accomplished by computing
* Numbers underlines (!_/) indicate references at the end of this
paper.
-------
Table 1
Light-Duty Vehicle Registrations by
Model Year and Calendar Year
(millions)
Model
Year
1977
1976
1975
1974
1973
1972
1971
1970
1969
1968
1967
1966
1965
1964
1963
1962
1961
Calendar Year
—
—
—
—
—
—
—
4.55
5.50
—
—
—
—
• —
—
5.29
6.60
5.54
—
— :
—
—
5.84
7.34
6.62
5.45
j
—
—
—
6.40
7.85
7.31
6.62
5.38
—
—
—
6.23
9.01
7.82
7.30
6.57
5.28
—
—
5.82
8.85
8.94
7.73
7.18
6.40
5.02
*™ ™"
6.18
8.12
8.83
8.93
7.66
7.05
6.18
4.65
6.45
8.92
8.05
8.79
8.85
7.53
6.82
5.80
4.08
9.92
6.28
8.81
7.87
8.53
8.50
7.11
6.26
5.05
3.26
5.92
9.28
8.88
8.80
7.77
8.31
8.17
6.65
5.62
4.27
2.52
7,16
8.91
9.12
8.85
8.59
7.49
7.93
7.58
4.92
4.71
3.34
1.82
7.98
10.15
8.71
8.88
8.61
8.29
7.12
7.33
6.71
4.96
3.69
2.47
1.26
6.43
11.26
10.14
8.62
8.61
8.49
7.93
6.62
6.53
5.71
3.97
2.82
1.81
0.90
4.68
9.76
11.33
10.09
8.54
8.34
8.33
7.55
6.11
4.79
4.82
3.23
2.22
1.40
0.68
6.47
7,68
9.74
11.13
9.87
8.24
7.96
7.77
6.85
5.36
4.88
3.92
2.57
1.74
1.08
0.52
7.17
9.55
7.47
9.59
10.85
9.56
7.86
7-**w
6.96 .
5.85
4.41
3.88
3.02
1.96
1.31
0.81
2.093*
* Registrations of all previous model years.
-------
Table 2
Average Ratio of Vehicles Surviving
from One Calendar Year to the Next
versus Vehicle Age
Years of
Vehicle Life Survival Ratio
1 to 2 1.386
2 to 3 1.022
3 to 4 0.991
4 to 5 0.981
5 to 6 0.972
6 to 7 0.959
7 to 8 0.933
8 to 9 0.895
9 to 10 0.856
10 to 11 0.813
11 to 12 0.785
12 to 13 0.773
13 to 14 0.773
14 to 15 0.763
15 to 16 0.755
All Years Beyond 16 0.750
-------
5.
vehicle survival ratios for each succeeding model year and projec-
ting the new 'vehicle sales for each required future model year.
First, to compute the survival ratios, the ratio of vehicles
surviving from each year to the subsequent year was computed for
each of the years of the life of the vehicle. For example, the
ratio of 1975 vehicles registered in 1975 to those 1975 vehicles
registered in 1976 were computed, as well as 1976 vehicles regis-
tered in 1976 and 1977, 1977 vehicles registered in 1977 and 1978,
etc. All of the ratios for survival from the first to second year
of vehicle life were averaged as were the ratios for each sequen-
tial year. These average survival ratios are given in Table 2, and
may be conveniently described as the survival probability vector.
It may be noted that the first elements of the survival
probability vector are greater than one. This may appear surpris-
ing, however it is a logical result of using registration data
compiled in June of each year. For example, assume x 1977 vehicles
were sold and registered prior to July 1, 1977, and that an addi-
tional y 1977 vehicles x^ere sold and registered after July 1, 1977.
Neglecting the destroyed vehicles, the ratio surviving until July
1, 1978 would be computed as (x + y)/x, which of course, is greater
than 1.0.
In general, a computation problem is incurred because the
vehicle model year begins in September or October, the calendar
y_e.ar starts. January 1, and the vehicle registration data are
reported as of July 1. Since the purpose of this report is to
present a simple method for investigating the relative aspects of
fuel economy programs, this 'problem is not treated in detail.
However, it is suggested that one approach to improve the predic-
tion system would be to research quarterly, or monthly, new vehicle
sales, and use this information to predict the vehicle use in the
first months after the vehicle is sold, but before it is recognized
in the July registration data.
The survival rate vector can be used to predict the number of
vehicles of any model year existing in a given calendar year, if
the number registered in any previous calendar year is known.
Consequently, the survival rate vector can be used to predict the
number of vehicles present in future years for all model year
vehicles since 1977. For model years later than 1977, which was
the last year of available registration data, some method must be
used to predict the new vehicle sales in each year.
Considering the available data, it was decided to indirectly
predict the new vehicle sales by predicting the total vehicle
population. This approach has several advantages; first when
attempting to model annual .fuel consumption, the total vehicle
population is a more important parameter than new vehicle sales.
Second, vehicle sales vary considerably with the state of the
-------
Table 3
>tal Light-Duty Vehicle Registrations
versus Calendar Year
Year Registrations
(in millions)
1965 68.9
1966 71.3
1967 73.0
1968 75.4
1969 78.5
1970 80.4
1971 83.1
1972 86.4
1973 89.8
1974 92.6
1975 95.2
1976 97.8
1977 99.9
1978* 102.8
1979* 105.5
1980* 108.2
1981* 110.9
1982* 113.5
1983* 116.2
1984* 118.9
1985* 121.6
* Predicted.
-------
7
national economy. Finally, predicting the total vehicle population
minimizes the tendency to accumuate errors which could occur if new
vehicle sales were predicted.
Table 3 gives the available total vehicle population data.2j
These data were fitted with a linear regression to predict total
future vehicle populations. These predicted future total vehicle
populations are also given in Table 3, while the existing data and
the predicted populations are plotted in Figure 1.
The survival ratios and the total population data provide the
necessary information to predict the vehicle population distribu-
tion. Starting with the 1977 model year, the most recent year in
which the population distribution is known, the survival ratios
were used to predict the population distributions for 1977 and
earlier model year vehicles in calendar year 1978. The sum of all
1978 and earlier model year vehicles was then computed and subtrac-
ted from the total predicted population for 1978 vehicles. The
difference between the total population and the vehicles existing
from previous years was assumed to be the 1978 model year sales.
This process was then iteratively repeated for all subsequent
years. The resulting predicted vehicle population matrix is given
in Table 4.
2. Annual Vehicle Miles Traveled
The next required parameter is the number of the average
annual vehicle miles traveled as a function of vehicle age.
Unfortunately, little recent data are available on this parameter.
The most frequently cited reference is the U.S. Census Bureau data
from 1970. V These data are 'presented in Table 5. It should be
noted that the .Census data indicate a very high number of miles
traveled during the first year the vehicle is in use. This implies
that there is an initial period of intensive use for new vehicles.
However, it might only reflect the manner in which some of the data
were obtained. For example if a vehicle which was purchased in
March had accumulated 10,000 miles by the time is was sampled in
August, just after a 5,000 miles summer vacation trip, this could
be interpreted as a vehicle accumulating distance at the rate of
20,000 miles per year. If vehicle use in the first few months
after purchase were more extensively researched this potential
problem could be reduced.
Since the Census Bureau data appeared to have some anomalies,
such as vehicles of some ages traveling farther than newer vehi-
cles, it was .decided to "smooth" the data by fitting it with a
linear regression. This approach had the additional advantage that
the regression could be used to predict the annual vehicle miles
traveled for vehicles more than ten years old. The vector of
predicted annual miles traveled, which was used in all subsequent
calculations, is presented in Table 6.
-------
Total Vehicle Registrations
versus
Calendar Year
Vehicle
Registrations
(millions)
120
100 .
80 -.
60 •
1965 1970
Predicted Registrations
ORegistration Data
1975
Year
1980 1985
FIGURE 1
-------
Table 4
Predicted Light-Duty Vehicle Registrations by
Model Year and Calendar Year
(millions)
Model
Year
1985
1984
1983
1982
1981
1980
1979
1978
1977
1976
1975
1974
1973
1972
1971
1970
1969
1968
1967
1966
1965
1964
1963
1962
1961
1960
1959
Calendar Year
1975*
4.68
9.76
11.33
10.09
8.54
8.34
8.33
7.55
6.11
5.79
4.82
3.23
2.22
1.40
0.68
0.52
1.74+
1976*
6.47
7.68
9.74
11.13
9.87
8.24 .
7.96
7.77
6.85
5.36
4.88
3.92
2.57
1.74
1.08
0.52
1.94+
1977*
7.17
9.55
7.47
9.59
10.85
9.56
,7.86
7.44
6.96
5.85
4.41
3.88
3.02
1.96
1.31
0.81
2.09+
1978;
8.26
9.95
9.77
7.41
9.41
10.55
9.17
7.33*
6.66
5.96
4.76
3.46
3.00
2.33
1.50
0.99
2.18+
1979
7.86
11.44
10.16
9.67
7.26
9.14
10.11
8.55
6.56
5.70
4.84
3.73
2.68
2.32
1.78
1.13
2.38+
1980
8.19
10.89
11.70
10.75
9.49
7.06
8.77
9.44
7.65
5.62
4.64
3.80
2.89
2.07
1.77
1.34
2.63+
1981
8.29
11.36
11.13
11.59
9.88
9.23
6.77
8.18
8.44
6.55
4.57
3.64
2.94
2.23
1.58
1.33
2.98+
1982
8.43
11.49
11.61
11.03
11.37
9.60
8.85
6.32
7.32
7.23
5.32
3.58
2.81
2.27
1.70
1.19
3.. 24+
1983
8.52
11.68
11.74
11.50
10.82
11.05
9.21
8.25
5.65
6.27
5.88
4.18
2.77
2.17
1.73
1.28
3.32+
1984
8.61
11.81
11.94
11.64
11.28
10.52
10.60
8.59
7.39
4.84
5.09
4.61
3.23
2.14
1.60
1.30
3.46+
1985
8.73
11.94
12.07
11.83
11.42
10.97
10.09
9.89
7.69
6.32
3.93
4.00
3.56
2.50
1.63
1.25
3.57+
* Actual registration data presented for these years.
+ Registrations for all previous years.
-------
10
Table 5
Average Annual Miles
versus
Vehicle Age
Year of Miles Traveled
Vehicle Life (thousands)
1 17.5
2 16.1
3 13.2
4 11.4
5 11.7
6 10.0
7 10.3
8 8.9
9 10.9
10 8.0
Table 6
Predicted Average Annual Miles
versus
Vehicle Age
Year of Miles Traveled
Vehicle Life (thousands)
1 15.9
2 14.9
3 14.0
4 13.1
5 12.2
6 11.3
7 10.4
8 9.5
9 8.6
10 7.7
11 6.8
12 5.9
13 5.0
14 4.0
15 3.1
16 2.2
17 and older 1.3
-------
11
3. Vehicle Fuel Consumption by Model Year
The EPA data are the best indication of average annual vehicle
fuel consumption. 4/ These data, for both the EPA city cycle and
the composite city/highway cycle, are plotted versus model year in
Figure 2. This plot shows that a significant change occurred in
fuel economy trends in 1975, the first year of the EPA voluntary
fuel economy program. Assuming current improvements in the fuel
economy of vehicles will continue, the future fuel economy trends
were predicted from a linear regression of the data since 1974.
These regression lines are shown in Figure 2, as are the current
fuel economy standards for 1978 through 1985. Since the regression
lines of the composite city/highway values are greater than the
current standards, while the city cycle values are less than the
standards, these regression predictions appear to be reasonable.
The fuel economy prediction model requires the fuel consump-
tion of the vehicles be known. For this reason fuel consumptions,
the reciprocal of the fuel economies, were calculated from the
available data, and are presented in Table 7. Also presented in
this table are the predicted fuel economy values and the subsequent
predicted fuel consumptions for 1979 and later model years. The
two fuel consumption columns of the table may be conveniently
considered as the fuel consumption vectors for the city cycle and
the composite cycle.
The vehicle population matrix, Table 4; the annual vehicle
miles traveled vector, Table 6; and the fuel consumption vectors,
Table 7; complete all of the information necessary to use the fuel
consumption model equation (1)..
C. Predictions of Annual Fuel Consumption
Using the model, equation (1), the total fuel consumed, TFCON.
for each of the years of interest can be computed from the vehicle
distribution matrix, VMIX^. given in Table 4, the vehicle miles
traveled, MIT;. given in Table 6, _and the fuel consumption, FC,
given in Table 7. The results of this calculation are presented in
Figure 3 and Table 8.
The computer program used to perform the fuel consumption
calculation is given in the attachment of this report. It should
be noted that this program divides the miles traveled during the
first year of vehicle life by 2 prior to multiplying by the number
of new model year vehicles. This is done because it is assumed
that these vehicles have, on the average, been in service for only
one half of the year prior to appearing on the annual registration
data list.
In 1975, it is estimated that 76.01 billion gallons of fuel
were consumed by passenger cars.5/ Based on this estimate, predic-
-------
12
Fuel Economy
versus
Vehicle Model Year
JU
25 -
20 •
Fuel
Economy
(mi/gal)
15 •
10 -
19
D EPA composite cycle fuel
economy results. /
OEPA city cycle fuel economy X^-
results. /
•fcFuel economy standards. / /
/* /
/v\
f / ^ — Predicted composite
•+/ cycle fuel economies
D ,7
D o \.
. o ^ — Predicted city cycle
DO . . . fuel economies.
D D D D D D O
D D
o o n n o
0 ° 0 0 o
°0
65 1970 1975 1980 1985
Year
Figure 2
-------
13
Year
Table 7
Average Fuel Economy and Fuel Consumption by
Model Year
City Fuel
Economy
(mi/gal)
Pre-1968
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979*
1980
1981
1982
1983
1984
1985
13.6
13.2
13.2
13.1
12.9
12.6
12.3
12.2
13.5
15.4
16.3
17.0
18.6
19.8
21.1
22.3
23.6
24.8
26.0
City Fuel
Consumption
(gal/mi)
0.074
0.076
0.076
0.076
0.078
0.079
0.081
0.082
0.074
0.065
0.061
0.059
0.054
0.050
0.047
0.045
0.042
0.040
0.038
Composite
City/Highway
Fuel Economy
(mi/gal)
15.8
15.4
15.4
15.
15.
15.0
14.5
14.4
15.6
17
18
19
21
22
23
25
26.6
27.9
29.2
Composite
City/Highway
Fuel Economy
(gal/mi)
0.063
0.065
0.065
0.065
0.066
0.067
0.069
0.069
0.064
0.056
0.054
0.051
0.047
0.044
0.042
0.040
0.038
0.036
0.034
* Values for model years past 1978 are predicted from a linear
regression of the data from 1975 through 1978 inclusive.
-------
14
National Annual Fuel Consumption
80
70. ..
60 ..
50 .-
Predicted fuel consumption
using EPA city cycle fuel
economy values.
DOT Data on
annual fuel
consumption.
Predicted fuel consumption
using EPA composite cycle
fuel economy values.
1970
1975
1980
Year
1985
FIGURE 3
-------
15
Table 8
Predicted Total Annual Fuel Consumption
Consumption Predicted Using
City Cycle Fuel Economies
Year (billions of gallons)
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
78.1
77.3
76.1
75.5
74.9
73.3
71.4
69.3
67.0
64.7
62.4
Consumption Predicted Using
Composite Cycle Values
(billions of gallons)
66.5
65.9
65.1
64.7
64.4
63.2
61.8
60.2
58.4
56.6
54.8
-------
16
tions using the EPA composite fuel economy values appear to under-
estimate actual fuel consumption by twelve to thirteen percent.
Using the urban cycle fuel economy values in the prediction model
appear to provide better accuracy. In this case, the model over-
estimates the annual fuel consumption by about three percent. This
is consistent with other observations, and with the decision to use
only the urban fuel economy results for vehicle labels, beginning
with the 1979 model year._6/ Using the urban fuel consumption, the
predicted values are in good agreement with the reported data.
III. Conclusion
The model predictions agree well with reported data for the
current years. It is therefore concluded that the prediction
accuracy of the model should be quite good since major changes in
vehicle usage are not expected in the next ten years. Even if
unanticipated changes do occur in vehicle use, predicted relative
effects of different technologies should still be valid. It is
therefore recommended that the model be used primarily for the
prediction of the relative effects of different technologies on the
annual fuel consumption.
-------
17
References
I/ H.H. Gould and A.C. Malliaris, "Highway Fuel Consumption
~~ Computer Model," Department of Transportation Report, DOT-TSC-
OST-73-43, April 1974.
2/ R.M. Lienert (editor), "Cars Still in U.S. Use by Year
Models," Automotive News, Detroit, Michigan, July 17, 1978.
3/ H.E. Strate, "Annual Miles of Automobile Travel," Nationwide
~~ Personnal Transportation Study, U.S. DOT, Report No. 2, 1972..
J.D. Murrell, "Light-Duty Automotive Fuel Economy ... . Trends
Through 1978," Society of Automotive Engineers, Paper No.
780036.
5/ W.F. Gay, National Transportation Statistics, Department of
~ Transportation Annual Report, DOT-TSC-OST-77-68, 1977.
6/ Federal Register, May 17, 1978 (43 FR 21412).
-------
18
Attachment
Fuel Consumption Prediction Program
-------
> . i c
> 2 C
> 3 C THUS PROGRAM IS DESIGNED TO CALCULATE THE ESTIMATED TOTAL VEHICLE MILES
> 4 C TRAVELLED PER YEAR BY PASSENGER VEHICLES AND TH.E ESTIMATED TOTAL GALLONS
> 5 C OF GASOLINE CONSUMED.
> 6 C
> 7 C
> Q DIMENSION VEHPOP<27» 11 > » VMILES<30 > r CITYFE<30) rVMT(30> rGASC<30> rSVMTC 11 > »SGASC< 11 >
> 9 C . • '
> 10 C
> 11 C THIS SEQUENCE PRESETS THE ELEMENTS OF THE ARKAYSr DESIGNATED SVMT AND
> 12 C SGASC RESPECTIVELY* TO BE ASSIGNED THE VALUE OF ZERO,
> 13 C
> 14 C
> 15 DATA SVMT/11*0./
> 16 . . DATA SGASC/11*0./
> 17 C
> 18 C
> 19 C THIS SEQUENCE READS IN THE PREDICTED MATRIX* DESIGNATED VEHPOPr THE
> 20 C ESTIMATED AVERAGE ANNUAL MILES PER AUTOMOBILE BY YEAR MODELr AND THE
> 21 C FUEL ECONOMY STANDARDS FOR PASSENGER VEHICLES,
> . 22 C
••> :.'3 c .
> 24 READ(5rlOOO)«VEHPOP 25 1000 FORMAT<12X,11FB.O)
> 26.2 2000 FORMAT
> ;:8 50 L-10
> 'J9 C
> 30 C . . .
> 31 C THIS STATEMENT WAS INTRODUCED WITH THE ASSUMPTION THAT A VEHICLE IS
> 22 C DRIVEN HALF THE ANNUAL VEHICLE MILES ITS FIR9T YEAR OF VEHICLE LIFE?
> 33 C SINCE THE REGISTRATION DATA WERE OBTAINED ON JULY 1 OF EACH YEAR.
> 34 C
> 35 C
> 36 VMILES 37 C
-------
> 38 C
> 39 C THE FOLLOWING TWO DO LOOPS SCAN EACH SUCCESSIVE COLUMNf ROW-DY-ROW
> 40 C AND PERFORM THE FOLLOWING OPERATIONS,
> • 41 C
> 42 C
> 43 DO 20 K=lfll
> 44 DO 10 J=lrl7
> 45 JJ=JH.
> 46 C
> 47 C
> 48 C THE FOLLOWING EQUATION CALCULATES THE TOTAL ANNUAL MILES TRAVELLED
> 49 C BY VEHICLES .OF EACH MODEL YEAR IN EACH CALENDAR
> 50 C
> 51 C
> 52 VMT(J)= 53 C
> 54 C
> 55 C THE FOLLOWING EQUAITON CALCULATES THE GALLONS PF§ GASOLINE CONSUMED
> .56 C BY VEHICLES OF EACH MODEL YEAR IN EACH CALENDAR YEAR,
> 57 C
> 50 C .
> 59 40 GASC 60 SVMTCK)=SVMT
> 61 SGASC=SGASC 62 10 CONTINUE
> 63 L=L-1
> 64 20 CONTINUE
> 65 C
> 66 C
> 67 C THIS SEQUENCE WRITES OUT THE TOTAL VEHICLE MILES.TRAVELLED AND THE
> 68 C TOTAL GALLONS OF GASOLINE CONSUMED PER YEAR,
> 69 C
> 70 C
> 71 WRITE(7f3000)
> 72 3000 FORMAT('!')
> 73 WRITE(7r4000>
> 74 4000 FORMAT<10Xr'TOTAL VEHICLE MILES TRAVELLED PER YEAR'»10X»'TOTAL GALLONS OF GASOLINE CONSUMED PER YEAR')
> 75 : URITE(7r5000)((SVMTrSGASC 76 5000 F.ORMAT('0'F23XrFll,3»40X»F9,3)
> 77 STOP
> 70 END
I
------- |