530SW142C
Prepublication issue for EPA libraries
and State Solid Waste Management Agencies
YOSEMITE NATIONAL PARK
BEVERAGE CONTAINER DEPOSIT EXPERIMENT
FINAL REPORT
This final report (SW-142a) describes work performed
for the Federal solid waste management program
under contract no. 68-01-2981
and is reproduced as received from the contractor
Copies will be available from the
National Technical Information Service
U.S. Department of Commerce
Springfield, Virginia 22161
A. jk .08812
U.S. ENVIRONMENTAL PROTECTION AGENCY
1977
-------
This report as submitted by the grantee or contractor has been
technically reviewed by the U.S. Environmental Protection Agency (EPA).
Publication does not signify that the contents necessarily reflect the
views and policies of EPA, nor does mention of commercial products
constitute endorsement or recommendation for use by the U.S. Government.
An environmental protection publication (SW-142c) in the solid waste
management series.
-------
ACKNOWLEDGMENT
This project was conducted by the Research Triangle Institute,
Research Triangle Park, North Carolina, pursuant to Contract No.
68-01-2981 with the Environmental Protection Agency. The statements,
findings, and conclusions presented in this report do not necessarily
reflect the views of the Environmental Protection Agency.
Tayler H. Bingham was project manager for this study. Jerome A.
Olson was primarily responsible for the development of the statistical
estimates. Juanita Daber assisted in writing this report.
Our appreciation is extended to individuals of the Yosemite Park
and Curry Co. and of the National Park Service for providing the neces-
sary data and interpretations. In particular, from Curry Co. we would
like to cite Edward Hardy, Chief Operating Officer; Garrett DeBell; Art
Goertzen, Jr.; John Crofut and William Letts and from National Park
Service, Leonard McKenzie and Jeri Andrews. Marion Thompson of EPA was
project officer for this study. Her interest and guidance are very much
appreciated.
in
-------
TABLE OF CONTENTS
Pjge
ACKNOWLEDGMENT Ill
LIST OF FIGURES vii
LIST OF TABLES ix
1. Introduction 1
2. Findings 2
3. Yosemite National Park 3
4. The Deposit System 5
5. Deposit System Impacts 6
5.1 Beverage Sales and Packaging Mix 6
5.1.1 Sales 7
5.1.2 Packaging Mix 13
5.2 Container Return Rates 15
5.2.1 Vended Sales 21
5.3 Recycling, Solid Waste, and Litter 22
5.4 Economics 23
5.4.1 Handling Costs 24
5.4.2 Revenue 27
5.4.3 Gross Profits 28
5.4.4 Net Profits 31
BIBLIOGRAPHY 35
APPENDIX
A: DAILY SALES AND RETURNS A-'!
B: MONTHLY VENDED SALES B-'l
IV
-------
LIST OF FIGURES
Figure Page
1. Daily beverage sales 8
2. Daily beverage sales and returns 17
3. Daily return rate 19
-------
LIST OF TABLES
Table Page
1. National Park System, 1974 4
2. Recycled beverage containers, summer 1976 23
3. Container handling costs by activity 25
4. Container scrap values 29
5. Gross profit rates by container and return rates ... 30
6. Total handling costs and revenues by activity .... 32
7. Summary of the economics of the Yosemite Deposit
System 33
-------
YOSEMITE NATIONAL PARK DEPOSIT EXPERIMENT
1 . Introduction
On May 17, 1976 the Yosemite Park and Curry Co. with the support
of the National Park Service and the Environmental Protection Agency
voluntarily instituted a 5-cent deposit on all beer and soft drink-
containers, both retail and vended, sold in Yosemite National Park,
California. Prior to the promulgation of Beverage Container Guidelines
applicable to Federal installations in September "'976, the Yosemite Park
and Curry Co. decided to make the deposit system a permanent operating
feature. The purpose of the deposit is to provide consumers with a
monetary incentive to return empty containers for a deposit refund.
This reduces the probability that containers will be discarded as litter.
Beverage containers are returned to the beverage distributor if they are
refiliable bottles or sold as scrap if they are nonrefillable bottles or
cans. The sale of returned containers as scrap helps to finance the
program. It also reduces the solid waste management burden in the Park,
Additionally, other environmental benefits are obtained when recycled
materials are substituted for virgin materials at the production level.
The purpose of this report is to present the results or impacts of
the deposit experiment over its first summer of operation.
-------
The major impacts examined in this report are the effects of the
Yosemite experimental deposit system on:
1) beverage sales and packaging mix
2) container return rates
3) recycling, solid waste, and litter
4) economics
The general findings of the experiment are summarized below and
discussed in detail in the remaining sections of this report.
2. F i ndi ng s
1) beverage sales and packaging mix - Two analyses of the impact of
the deposit on beverage sales were conducted, one based on total sales
and one on vending machine sales. Both analyses indicated that 1976
beverage sales, both retail and vended, were below the sales expected for
1976. However, it is not possible to present convincing evidence as to
the cause of the sales declines since factors other than deposits (e.g.
weather, number of visitors to the Park, prices) that are known to affect
beverage sales are difficult to adjust for with the available data.
In mid-summer beverages in bimetal containers (7-Up and Shasta) were
discontinued due to the absence of an acceptable secondary market for the
returned bimetal cans. With this exception consumers apparently con-
tinued to purchase beverages in the available alternative brand/container
combinations in about the same proportions as they had before the deposit
system was initiated. Ninety-eight percent of carbonated beverage sales
in Yosemite are in cans.
2) container return rates - Approximately 7 out of 10 beverage
containers sold in Yosemite during the summer were returned.
-------
3) recycling, solid waste, and litter - About 30 tons of beverage
containers were recycled. Of this total 74 percent was aluminun; 15
percent, steel; and 11 percent, glass. Approximately 83 percent of the
total recycled beverage containers had been returned for the deposit
refund. Solid waste was also reduced by«30 tons. Cans marked for
deposit were seldom found as litter in the Park.
4) economics - The economics of the deposit experiment were
estimated by comparing revenues with costs. Revenue is available from
the scrap value of the returned containers and the forfeited deposits on
containers which are not returned. Costs include handling (labeling,
collecting deposits, returning deposits, disposing of containers and
bookkeeping), equipment and supplies, and managerial and supervisory
time. Considering all costs Curry Co. estimates that the system broke
even in the first summer of operation. Eliminating costs which were
unique to the first summer of operation, the system would yield an
expected profit of approximately $8,000 in future summers. Profit
increases to approximately $12,000, if one further subtracts the Yosemite
Park and Curry Co. estimate of the value of supervision and management
time which they would expect to allocate once the system was in
operati on.
3. Yosemite National Park
Yosemite is one of 38 National Parks in the system controlled by the
National Park Service (NPS). It has more overnight visitors than any
other park (table 1). Yosemite is located in eastern California, about
218 miles from San Francisco and 275 miles from Los Angeles in the heart
of the Sierra Nevada Range. The Park, which is 1,189 square miles in
-------
Table 1. National Park System, 1974
Number
of areas
National Park System 309
National parks 38
Major national parks
Great Smokey Mountai ns, North
Carol i na-Tennessee
Grand Teton, Wyoming
Acadia, Maine
Olympic, Washington
3ocky Mountain, Colorado
Hot Springs, Arkansas
S hen an do ah, Virginia
Yosemite, California
Vei lows tone, Wyoini ng- Idaho-
Montana
Grand Canyon, Arizona
Mammoth Cave, Kentucky
Hawaii Volcanoes, Hawaii
Glacier, Montana
Mount Ranier, Washington
Zion, Utah
National monuments 83
National parkways 6
National recreation areas 16
National seashores 9
National capital parks 1
Other 156
National historical, archeo-
logical, and military areas
Miscellaneous other areas
Annual vis
Total
217.4
53.1
38.4
9.6
3.0
2.8
2.7
/ , 0
2.4
2.4
2.3
2.0
"i .9
1.8
1.4
1.4
1.2
1.0
15.2
34.5
18.1
11.5
8.7
75.9
72.9
3.0
its (millions)
Overnight
15.4
NA
7.5
0.6
0.5
0.2
0.5
0.3
<0. 1
0.6
1.8
1. 1
0.8
0.2
0.1
0.4
0.2
0.2
NA
NA
NA
NA
NA
NA
NA
NA
NA - not applicable
Source: U.S. Department of the Interior, National Park Service, 1974,
-------
area or about the size of the State of Rhode Island, is administered by
the National Park Service. Most concessions in the Park including all
beverage concessions are operated for the NPS by the Yosemite Park and
Curry Co. (Curry Co.), which in turn is owned by the MCA, Inc.
For the last several years the Park has received more than two mil-
lion visitors annually. About 80 percent visit Yosemite during the
months from May through September. This experiment ran for 126 days,
from May 17 to September 19, 1976. During this time there were about
1.3 million visitors. For the month of August, which is the peak month
of the year, there was an average of about 15,000 people (both visitors
and employees) in the Park each day--about 12,000 staying overnight.
Between 70 and 80 percent of all visitors come to Yosemite from
California and nearly 20 percent from Los Angeles County alone. Out-
of-state visitors primarily come from western, mid-western and middle
Atlantic states. The largest proportion of the visitors (44 percent)
are between 18 and 29 years of age and the largest occupational cate-
gory represented is professional and technical workers (40 percent).
Over 70 percent of the visitors have attended college and nearly one-
half have a college degree. Over 70 percent of those visiting Yosemite
stay no more than two nights (U.S. Department of Interior, 1973).
4. The Deposit System
Yosemite has 23 retail outlets and 34 vending machines dispersed
over the entire Park area. Approximately 80 percent of beverage sales
are through the 19 retail outlets and 30 vending machines in the
-------
Yosemite Valley. Soft drinks and beer are delivered directly to the Tare
retail outlets, otherwise to a central warehouse or, in the case of
beverages for vending machines, to a storage trailer.
The containers are then hand-stamped to indicate that a deposit is
to be paid on the container. The stamping is done either at the retail
outlet, or in the case of vending machine drinks, at a separate storage
area. Deposits are collected at the poi nt-of-sal e by the cashier or
incorporated as part of the sale price for vendi ng machi ne drinks.
Deposits are refunded upon presentation of the labeled container at any
retail establishment or at the central recycling center that is located
in the main shopping area of the Valley. Returned containers, other
than refill able bottles, are sorted by material, bagged, transferred to
a centrally located trailer in the Park, and stored, ,-Jhtn the trailer
is full it is taken to Fresno, California where the containers are sold
for scrap. Refillable containers are picked up by the distributor for
reuse.
5. Deposit System Impacts
The major impacts of the deposit system discussed below are: (1)
beverage sales and packagi ng mix, 2) container return rates, 3) re-
cycling, solid waste, and litter, and 4) economics.
5.1 Beverage Sales and Packaging Mix
The possible impacts a deposit system could have on the total sales
of beer and soft drinks and on the packaging mix have become major issue-
It is believed by some that a deposit system reduces beverage sales due
the increased out-of-pocket cost to consumers who do not return empty
-------
containers for a refund or the indirect cost associated with the incon-
venience of returning them.
Consumers who previously purchased beverages in nonref 111 abl e bottles
could shift e^the^ to refill able bottles (returning or discarding them) o>~
to cdns; consumers who previously purchased beverages in cans could shift
to >-ef il 1 abl e bottles.
The experiment did not provide the data needed to thoroughly evaluate
thf' s .jl es an') packaging mix ^n a statistically rigorous fashion.* A dis-
cussion of the possible change in beverage sales and the shift in con-
tai npr mix due fo the Yosemite experiment follows.
?.!.! Sales
Over the summer season 54,325 cases (at 24 containers per case) of
b'_-er and soft drinks were sold in the Park. The greatest majority of these
sales was in the Valley. According to Curry Co. personnel daily beverage
sales vary significantly (figure 1) for the following reasons:
i) the daily variation in the number of people in the Park, and
2; the variation in weather (perhaps the most significant
factor affecting beverage sales).
*To estimate the impact of deposits on sales and packaging mix ideal-
ly requires beverage sales and packaging data for previous periods, iden-
tification of independent variables other than the deposit (i.e. weather
conditions, price, income of visitors, type of container, available brands)
which may affect sales and packaging and historical estimates of these
variables. Given these data one would then statistically evaluate the ef-
fect of the independent variables on sales and the packagi ng mix. After
allowing for the effects of these variables in 1976, next examine the
extent that the predicted 1976 total sales and packaging mix differed from
the actual. The differences would be attributed to the deposit system.
-------
. . . . . .- . i .
6 8 10 12 If 16 18 ?,C 22 2f 26 28 30
June
(A/iA : I . !.. r. i I i I L..I.J !
2 f 6 8 10 12 If 16 18 20 22 21- 26 28 30
Legend:
. weekend
//// holiday
Figure 1. Daily beverage sales
july
-------
CO
r>
o
o
a:
.
or
c±
cr
a
en
a
cr
CO
O
CO
z
o
o
cr
CO
or
ac
a
2 1 6 8 10 12
16 18 2C 22 21 26 28 30
DRY
Aug.
20
18
16
If
12
10
8
6
' ' '
I
2 1 6 8 10 12 11 16 18 20 22 21 26 28 30
DRY Legend:
Sept. _ : i weekend
'//// holiday
Figure 1. Daily beverage sales (Cont.)
-------
Thus one could hypothesize that per capita beverage sales were rela
to weather.* Estimates of this relationship were statistically develope
*To test for the effect of price variation on beverage sales a pric
variation must exist. Since there was no price variation in the Park
during the 1976 period this effect could not be estimated.
**The procedure was as follows. First, estimates of the number of
people in the Park daily during the peak season were developed from NFS
data. Daily weather data for the Park were obtained from the U.S.
Department of Commerce. Finally, statistical relationships between per
capita beverage sales and the weather conditions in the Park were
estimated.
The NPS counts the number of vehicles entering the Park daily. Eac
vehicle is estimated to average 3.5 visitors. Based on the point of ent
to the Park and survey estimates of the average length of stay of vi si tc
for each point of entry, estimates of the cotal visitor-hours can b>3 de-
veloped for each day of the summer. (NPS subtracts from the vehicle dat
NPS and other Park employee vehicles and also adjusts for bus loads of
visitors. However, this has only a small effect on their estimate of
visitor-hours.) Using the NPS vehicle data and the procedure outlined a
the total visitor-days for the 126-day period were e:"i^ated to be 2.675
million (see appendix A). In addition, there were 1,500 to 2,200 em-
ployees of NPS and Curry Co. in the Park during the peak season. Using
the 1,500 value raises the estimated total to 2.865 million person-days.
Since the only Yosemi te weather data avai 1 abl e were daily maximum ;
minimum temperatures for the Valley, the following dai 1y consumption mo(
was estimated for the 126 days using the standard stati sti cal technique
ordinary 1 east-squares regression:
(1 ) In PCS = 1.93 In TEN> - 8.98
(8.57) (-9.46)
R2 =0.40
where
PCS = per capita sales, units per person-day,
TENP = average temperature, F°, and
the number in parenthesis below the estimated coefficients is the t-stal
Logarithmic coefficients provide the best quantitative approach si
the coefficient on average temperature can be directly interpreted as t
"elasticity" of per capita sales to the average temperature.
This estimated model attributes 40 percent (R2 = 0.40) of the dail
variance in per capita beverage sales to the daily variance in the aver
temperature. The model's explanatory power is low due to the likelihoo
that the procedure used to estimate visitor-days is not as accurate as
should be. Also the lack of inclusion of other possible explanatory
variables besides weather has affected the results.
10
-------
The results of this analysis indicate that beverage sales in the
'ark werac;>'d 0.5 units dsily per capita.* The hypothesized effect of
weather on , al es .5 stati stically verifiedevery one percent increase
in average temperature "esults in an increase of 1.9 percent in per
capita sales. The convincing evidence of the effect of weather (i.e.,,
tenperature) on beverage sales emphasizes the desi rabi 1 ity of correcting
for this effect when evaluating the impact of deposits on per capita
s a Ies.
Two analyses of the sales impact of the deposit system were
conducted. In tK- first, equation (1) above wac solved for expected
dail , per api ta ales ising 1975 daily average ; emperatures over the
sane 126-cLy period as the deposit system was in existence in 1976.
Then, using the sane procedure as described above to estimate the number
of people in the Park daily, expected total daily sales for 1975 were
computed and summed over the 126-day period. This expected value was
then compared to the estimated actual sales for the 126-day period in
1975. "ihese results are shown below.
Beverage sales (cases)
Actual Estimated Expected
1976
entire year 73,953
126-day perid 54,325
1975
entire year 81,560
126-day period 59,913 56,952
*An unknown quantity of beer and soft drinks is brought into the
Park by vi si tors.
11
-------
The estimated sales value (59,913) was developed by applying the
126-day share of total 1976 beverage sales to the total 1975 beverage
sales value supplied by Curry Co. The results suggest that the expected
1975 sales (56,952 with deposit) for the 126 day period would have been
5 percent less than the estimated 1975 sales (59,913 without deposit)
*
for this period. We cannot, however, be certain about what to attribute
the 5 percent difference because other factors which could affect, per
capita sales were not controlled. Although this approach was an attempt
at measuring the effect of the Deposit on sales, it cannot be vigorously
defended as indicating that the deposit caused a sales decline in 1976.
A second analysis was conducted using vended soft drink sales for
the 1973-76 period. Vended sales are of particular interest because it
is frequently speculated that deposits would significantly reduce vended
sales since consumers of vended soft drinks are alleged to be especially
convenience oriented. During the vended test period, vended sales
represented about 13 percent of total beverage sales on a units basis.
Another consumption model was estimated using monthly vended sales
data, 1973-76, as presented in Appendix B. Because the price of vended
soft drinks rose from 25 to 30 cents per can beginning in August 1977 ;;t
was possible to include price variation in the model.* However, because
the number of observations is small we can expect to identify only
strong effects.
*The estimated monthly vended consumption model is:
(2) In PVVS = - 0.04 In YEAR + 0.33 In PRICE + 1.49 In TEMP - 7.87
(-0.51) (0.46) (2.97) (-3.01)
R2 = 0.39
Where
PVVS = per visitor vended sales, units per visitor,
YEAR = a dummy variable, set equal to zero for ye^rs
prior to 1976, set equal to one for 1976,
TEMP = average temperature, F°, and
PRICE - soft drink price, cents per unit divided by the
consumer price index.
-------
This model attributes 39 percent of the monthly variation in vended
s jf t drink sales per visitor to the combined effects of YEAR, PRICE, and
"ENP . Obviously the model is limited by the use of monthly data whereas we
v/ould have preferred to use daily data. It is interesting to note, however,
Hiat the coefficient (1.49) on average temperature for the monthly vended
node1 is similar to the temperature coefficient (1.93) in the daily model
and is stati sti cal 1 y si gnif icant at normal ly accepted confidence ''evels (95
percent). The price of vended soft drinks has a small positive effect on
vended sal es.*
The sign on YEAR is negative as is expected. Letting YEAR entirely
represent the effect of the deposit system, the best estimate of the true
effect of the deposit system shows a reduction in vended sales per visitor
of 4 percent. However with only 20 observations, 5 for 1976, the statisti-
cal quality of this estimate is poor. Specifically, 95 percent of the time,
with data of the quality used in this analysis, we would expect the true
value of vended sales per visitor to range between -20 percent and 0 percent
(with positive values being rejected on the belief that there is no reason
to expect a deposit system to increase vended sales).
5.1.2 Packaging Mix
At Yosemite over 98 percent of all the beverages are sold in cans.
Cans with nondetachabl e closures are preferred because the detachable
flip-tops are frequently littered; however, beverages are sold in cans
with and without detachable tops. For many years it has been the policy,
*Thi s peculiar result is not alarming since the t-statistic of 0.46
indicates that the estimated positive effect is so weak that it could be
obtained merely by chance.
13
-------
as requested by the NPS, to minimize sales of bottled beverages because of
the problems to humans and hazards to animals when bottles are littered and
broken. Some soft drinks are sold in large, mostly refillable bottles.
Bottl ed mixers, 10-ounce Schweppes products, and 12-ounce bottles of
imported beer are also sold in nonref il 1 aSl e bottles. In total, about
24,000 bottled beverages were sold in the Park in the 1976 summer season.
There was no change in the kinds of containers sold in the Park in 1976.
A deposit system may create the potential for shifts in the packaging
mix as consumers who discard the empty container would pay a higher price;
that is, they would forfeit the deposit and those who return the container
would incur some inconvenience. The result would be reflected in changes
in the rate of beverage sales by container typesome beverage-contai ner
combinations being purchased more frequently and others less frequently
than prior to the institution of the deposit system.* This behavior would
result in shortages for some beverage-contai ner combinations and surpluses
for others. Retailers would presumably order more of the products in short
supply and less of those for which surplus supplies existed experimenting
until they again found the new optimum order rate for each beverage-
container combination.
As noted above, Curry Co. is the prime concessionaire at Yosemite, It
offers for sale and sells primarily beverages in cans with prices being set
*The effect of brand loyalty and consumer choice is important to em-
phasize here. If the brand is offered in alternative containers then con-
sumers have the opportunity of switching containers without switching
brands. On the other hand if brands tend to be offered in only one con-
tainer type then the consumer must switch brands to switch containers.
Further, even though beverage distributors may offer their product in al-
ternative containers, retailers may not elect to stock all types.
14
-------
by the NFS. No change in the brand-contai ner combinations offered for sale
A as made by Curry Co. with one important exception discussed below. The
NPS policy of minimizing bottled beverage sales was not changed after insti-
tuting the deposit system.
In the absence of a competitive market with a broad choice of brands
and containers it is difficult to generalize regarding the impact of the
deposit system on package mix. However, in Yosemite no shortages occurred
t n t he refillables offered for sale nor were there any requests by visitors
for brands or sizes of beer and soft drinks in refill able bottles other than
those carried. Therefore, at least in this setting, no significant changes
in packaging mix were either possible or requested by visitors.
The exception noted above was the bimetal can. The closest market for
returned bimetal cans is South San Francisco whereas aluminum cans could be
sold to a scrap dealer in Fresno. Therefore, for all but a couple of
specialty beverages, those in bimetal cans were discontinued about mid-
summer. The primary brands affected were 7-Up and Shasta.
5.2 Container Return Rates
For the 126-day test Curry Co. personnel tabulated daily deposits paid
and returned for each of 23 retail and 34 vending locations. These data,
which are summarized in appendix A are the basis of the conclusions drawn
here.
Over the entire 126-day period 1,303,800 bottles and cans of beer and
soft drinks were sold in the Park. Deposits were refunded on 903,800 con-
tainers; therefore, the return rate for the summer averaged 0.69 or said
another way, almost seven out of every ten beverage containers sold were
returned. According to Curry Co. personnel operating the recycling center,
15
-------
probably a small share of the total returns are by scavengers who collect
littered containers or who obtain them from trash receptacles, dumpsters,
transfer stations and the landfill.
Daily beverage sales and returns are shown in figure 2, and simple
daily return rates (developed by dividing returns for the day by sales
for the day) in figure 3. The lagged structure of container returns,
and the effect of the day-of-week and day-of-summer on return rates were
i nvesti gated.
Briefly summarized, it was estimated that about 39 percent of con-
tainers sold are returned on the day of sale, 15 percent one day after,
6 percent wo day, aftf, 4 percent three days after and the remaining 5
percent mo e than three days after the date of sale.*
The day of the week analysis was confounded by the presence of an
anomaly in the reporting of the data.** Therefore, while only tentative
conclusions can be drawn, it does appear that Saturday return rates are
somewhat lower than average and Sunday somewhat higher. This would appear
to be confirmed by Curry Co. personnel at the recycling center who report
that Sundays are especially busy since for many visitors it is the day they
*To develop estimates of the time-di stri buti on of container returns
the following mode" was estimated using ordinary least squares regression.
(3) Rt = St + otiSt.! + «2St-2 + + anSt-n
where R = returns of containers on day t, and
S = sales of beverages on day t.
The R^ statistic for the model is 0.58. The coefficients
(a]_, «2' an) preceding the sales terms can be directly interpreted
as the share ot sales for that time (i.e., day) that were returned in
day t.
**Specifically, for several Thursdays vended sales were reported as
zero.
16
-------
CO
cc
CO
CO
z
o
o
cc
CO
a:
A = sales
+ = returns
°r 6 8 10 12
16 18 20 22 2°! 25 28 30
DRY
June
CO
a
-------
CO
a
cc
CO
CO
o
en
CO
ce
>-
en
a
= sales
+ = returns
10 12
f 16 18 20 2 ' 2^ 26 28 30
DflY
Aug.
CO
a
2:
~
' ' ' f ' ' ' I
A * sales
+ = returns
V r I i I r i i I i I i i t. 1
2 f 6 8 10 12
16 18 20 22 2f 26 28 30
DflY Legend:
Sept.
weekend
holiday
Figure 2. Daily beverage sales and returns (Cont.)
-------
tr
cc
UJ
Q±
>-
cz
c.
1 ' i ' I ' I ' [ ' I ' I ' I ' I'1 I ' I ' f 4 I ' I '
2 «r 6 8 10 12 If 16 18 20 22 24 26 28 30
June
a:
a:
LU
cc
i i i[ jIi|iIi r ii i ~iIii iIIilI
2 4 6 8 10 12 H 16 18 20 22 2^ 26 28 30
DflY Legend:
July
weekend
holiday
Figure 3. Daily return rate
19
-------
OI
ca
z.
C£
(XL
>-
cc
C3
5 8 10 12
16 18 2'J 22 2f 26 28 30
DfiY
Aug.
cr
o
2 °r 6 8 10 12 If 16 18 20 22 24 26 28 30
DflY Legend:
Sept.
vnekend
hoiidav
Figure 3. Daily return rate (Cont.)
20
-------
eave the Park. A test for a day-of-the sunnier effect indicated no signifi-
ant trend over the simmer in return rates.
5.2.1 Vended Sales
At the request of EPA Curry Co. personnel ran a special test on vended
ales to determine whether return rates from vended soft drinks varied sig-
if icant iy f rom those for nonvended beverages. Curry Co. personnel marked
'" /ended cans with a different color stamp during a one-week period, July
i to July 25. At the end of each day, 20 cans were sampled from each bag
r ^eturned cans. There are approximately 400 cans per bag. This sampling
as continued one week beyond the last day of the special marking of vended
>ntainers.
The expected proportions of container returns which should be represen-
;d by vended sales, based on the proportion of vended to total sales, were
>mpared to the actual proportions found in the sample. Based on the appl i-
ition of standard statistical techniques, it was concluded that there is
) strong evidence that the return rate of vended containers differs from
ie ^eturn rate of 69 percent for all containers.*
*The expected proportions were estimated from the time distribu-
on of container returns and the vended salei data. They were compared
> the actual proportions found in the sample and a chi-square test was
,ed to test for a significant difference.
This test statistic is computed as :
Z
(observed i - expected i)2
expected
lere "observed" is the proportion of specially marked cans found in the
mple of returned cans taken each day, and "expected" is the proportion
cans we could expect to find in the sample if the null hypothesis is
-Lie. "Expected" was computed as a distributed lag of past sales divi-
d by returns for the day, where the lag weights were those estimated
>r beverage returns as a whole. The statistic is distributed as chi-
juare with n-1 degrees of freedom.
21
-------
5. 3 Recycling, _SoT_id Waste, and Litter
In 1975 Curry Co. established a recycling program "...to reduce the
waste of material resources and energy, and to reduce to a minimum the n<
for garbage to be hauled to a land-fill" (Yosemite Park and Curry Company
1975, p. 2). A centrally located Recycling Center is maintained adjacent
to the Village Store and is open daily during the peak season. Additione-
centers are in employee housing areas. In 1975, 95 tons of waste materi;
were recycled including almost one ton of aluminum, one Ion of steel (bi-
metal), two tons of tin, 11 tons of glass ar;d SO tons of cardboard (Yosen
Park and Curry Company, 1975, p. 3). Other recycled materials include:
newspapers, copper, batteries, and tires. As a result of these efforts,
prior to the institution of the deposit system, Curry Co. had already i d<
tified markets for the materials present in returned containers and had
developed systems for handling returned materials.
During the 126-day test period, Curry Co. recycled about 30 tons of
glass, bimetal, and aluminum beverage containers. As shown in table 2, '
predominate material recycled was aluminum. Of the 30 tons, at least 83
percent represents containers which were returned for the 5-cent refund.
The deposit system apparently induced "piggy backing" as consumers also
returned containers not purchased in the Park. The returned containers \
bagged and shipped to Fresno where they were sold for scrap.
Obviously, the recycling of returned containers not only produces n
enue to help finance the deposit system but also reduces the quantity of
solid waste which otherwise must be collected, transported out of the Par
and properly disposed of in a landfill.
22
-------
Table 2. Recycled beverage containers, simmer 1976
Pounds
With
Material deposits
Glass NA
Bimetal 7,141
Aluminum 41,943
Total
Without
deposits
NA
1,548
2,173
Total
6,661
8,689
V
44,116~
59,466
NA - not available
_!/ Of this total collected, 37,025 pounds were
allowed by the recycling company.
Source: Yosemite Park and Curry Co.
Prior to the institution of the deposit system there were no sur-
veys of litter in Yosemite which could serve as a benchmark for estima-
ting the impact of the deposit system on container littering. However,
NPS personnel responsible for collecting litter believe that fewer lit-
tered containers were collected in the summer of 1976 than in previous
years. Reportedly, refund bearing containers were seldom found.
5.4 Economics
The cost of the deposit system to a Park concessionaire is estimated
by subtracting the cost of labeling, handling, bookkeeping and other rele-
vant costs (including the time and cost of designing the system, designing
and printing consumer information material, running special tabulations,
and meeting with and answering questions from EPA, RTI, the press and other
interested parties) from the forfeited deposits on containers not returned
and the scrap value received when returned containers are recycled. Of
23
-------
special mention is the treatment of solid waste costs. If recycled
containers result in a reduction in solid waste cost then a credit for
these cost savings should also be included.
Based on a simple model of the costs of a deposit system, presented
below and more fully discussed in A Guide to Implementing a Deposit-
System (Youngblood and Bingharn, 1977), estimates of the cost of the
system at Yosemite were calculated. In developing the cost estimates
presented here the approach was to first present those costs of the
deposit systan which primarily relate to beverage sales or container
returns. Then various other- costs were included which tend to be re-
lated to the existence of the deposit system rather than to the volume of
beverage sales or container returns. To the extent possible the results
were generalized so that the reader may visualize the economics of a
deposit systan in other situations.
Although the values were, of course, developed from the Yosemite
experiment, what is presented should not be interpreted as representing
a rigorous analysis of the Curry Co. costs.
5.4.1 Handling Costs
The total handling costs of a deposit system include the extra ex-
penditures to:
1) label the container,
2) collect the deposit,
3) return the deposit,
4) dispose of the containers, and
5) keep records of costs and revenues.
These five categories of expenditures in addition to the associated
labor costs are discussed below and summarized in table 3.
24
-------
Table 3. Container handling costs by activity
1)
2)
3)
4)
5)
Activity
Label
Collect, deposit
Return deposi t
Dispose of containers
Account! ng
Productivity
(man hours per
100 cases)
I/
1.00
21
0.23"
21
1.86"
3/
4.87~
4/
0.12"
Labor cost
(dollars per
manhour)5
$ 3.89
3.89
3.89
3.34
4.23
Activity cost
(dollars per
100 cases)
$ 3.89
0.89
7.24
6/
16.46"
0.51
J_/ Based on Yosemite experience with hand-stamping of 100 cases per manhour.
2_l From Alpha Beta Acme Markets, Inc.
3/ From Alpha Beta Acme Markets, Inc. for the following activities: 1) moving
full rack of bottles to storage area. 2) clearing of bottle storage area,
3) sorting of bottles from rack, 4) moving empty rack to front, and
5) checking empties to driver.
_4/ Based on Yosemite experience of 30 minutes daily.
5/ Based on Yosemite wages plus 36.5 percent fringe benefit factor.
6_/ Includes a mileage charge on the truck used for daily pickups of $0.19 per
100 cases.
25
-------
1) 1 abel- All containers at the Park were stamped by hand with a
soft sponge stamp. The cost of 24 stamps and 12 pints of frozen food
ink, which was chosen because it can withstand moisture, was $267. The
containers were stamped on the bottom at an estimated rate of 100 cases
per man hour.
2) col lect deposits - The salves clerk must expend additional time
with each beverage customer to ring up and collect the deposit. No
estimate of these costs is available for the Yosemite experiment; however,
a time and motion study indicates about. 0.3 seconds are required per
container to ring up the deposit. This value has been used here. On a
total basis, this would imply the equivalent of one additional employee at
Yosemite over the 126-day test period. Obviously, however, the total time
required to collect the deposit is distributed over all retail clerks and
is not likely to result in an employment increase.
3) return deposits - For all returned containers, the cashier must
ring up the deposit being returned and return the deposit to the customer.
No estimates of the time or cost of this activity have been developed by
Curry Co. However, based on the time and motion study, the time required
can be as high as 2.8 seconds per container. This value would imply the
equivalent of an additional 0.7 employees for the 126-day period.
4) dispose of containers - The returned containers are bagged, stored
and, when a truck load is accumulated, shipped to Fresno where they are
sol d for scrap.
At Yosemite the recycling center is manned from 9 a.m.-8.p.m. seven
days a week during the peak season. About four additional manhours are re-
quired daily to collect returned containers from the retail outlets and to
26
-------
truck the recycled materials to a central storage area prior to being
shipped to Fresno for a total of 15 manhours daily. Estimates of the
time spent at the retail outlets are not available. Assigning the total
of 15 manhours daily to the deposit system results in a productivity
rate (manhours per returned container) quite similar to the estimate
developed from a supermarket time and motion study.
5) accounting - The additional bookkeeping requirement is informally
estimated by Curry Co. to be 30 minutes daily or about 0.? seconds per
container on average for the summer season. This value is about equal
that of the highest value reported in the time and motion study.
6) labor costs - The cost of performing the above tasks is the
product of the productivity of the employee (manhours per container)
and the cost of the employee (dollars per manhour). These costs include
not only the wages paid the employee but also taxes and fringe benefits
which include lodging and meals. Curry Co. estimates these costs to
average 36.5 percent of the wage base. The recycling staff is paid
$2.45 per hour, sales clerks $2.85 and bookkeepers $3.10. The resulting
costs per 100 cases are shown in the fourth column of table 3.
5.4.2 Revenue
The primary revenues which accrue to retailers under this deposit
system are: 1) forfeited deposits, and 2) sales of the returned con-
tainers for recycling. Because the NPS charges Curry Co. per load of
collected solid waste and because frequency of collection has not been
altered, no solid waste costs were avoided.
27
-------
The scrap prices paid to the Curry Co. are shown in table 4. Since n
transportation costs were incurred by Curry Co. the FOB and delivered sera
val ue are the same.
5.4.3 Gross Profits
The average profits per container due to the deposit are
(4) P = R - C
where
P = profits per unit of beverage sold,
R = revenues per unit of beverage sold, and
C = cost per unit of beverage sold.
The costs can be divided into those costs incurred on every sale
plus those incurred when a container is returned or:
(5) C = (LABEL) + (COLL) + (ACC) + RR [(DISP) + (RET) ]
where
LABEL = cost of label ing,
COLL = cost of collecting deposit,
RET = cost of returning deposit,
ACC = cost of bookkeeping,
RR = fraction of containers returned, and
DISP = cost of disposing the returned container.
The revenues can be divided into the forfeited deposits and the
scrap sales of returned containers:
(6) R = (1-RR)(D) + (RR)(SV)
where
D = deposi t, and
SV = scrap value of the container.
28
-------
Table 4. Container scrap values
Container
Aliminum can
Steel can (bimetal )
Nonref il 1 abl e bottl e
beer (12 oz)
soft dri nk (1 6 oz)
Wei ght
(1 b per
100 cases)
105.6
264.0
919.2
1500.0
Scrap
(dollars
per ton)
$ 300
20
20
20
value
(dollars per
100 cases)
$ 15.84
2.64
9.19
15.00
Net scrap value
(dollars per
100 cases)
I/
$ 15.74"
2/
* ~~
11
6.02"
11
11.83
!_/ Calculated from transportation cost data provided by Curry Co.
2/ Not computed since nearest market is San Francisco, an uneconomi-
cal dTstance to ship scrap steel from Yosemite.
The gross profits thus depend on the return rate and the scrap value of
the container returned. It is necessary to examine each container because
each has a unique salvage value.
The resulting gross profit rates (dollars per 100 cases) for beverages
marketed in aluminum cans and glass bottles are:
aluninun can = 114.71 - RR(127.96),
nonref ill able beer bottle (12 oz) - 114.71 - RR(137.68), and
nonrefillable soft drink bottle (16 oz) = 114.71 - RR(131.87).
The above gross profits for the three container types at the Yosemite
return rate of 69 percent as well as a set of alternative return rates are de-
picted in table 5. This table clearly shows the inverse relationship between
the return rate and gross profits as the return rate goes up, profit decreases.
However, it should be pointed out that a major cost element is labeling. At
Yosemite this was done by hand at an estimated cost of $3.89 per 100 cases
(table 3) or approximately $2,000 for 54,000 cases. If containers can be
29
-------
*
O)
en
T3
CL)
CL)
<4-
o
LO
QJ
-
4
i
rc
a/
c
5
c
o
.a
(/
CD
£
-I-J
n_
o
OL
to
O
5-
O
.
LO
CU
r
r"l
fO
I
I/I
4-J
1°
£
-l~>
O
o
r
^ ^
r^ cn PJ O CTI IcTi
] «!.
LO CO f*~> LO OO fM
i CM * C\J ICNJ
1 1
CD
cn
O
r~~- CM to CM CM JO
( (O ^ *^ LO *^J
... . j .
LO LO O ^ LO iCTl
C^l CM 1 i CM I '
i .^ ' *-*
C
'CO
a: j o
rr
cz
r__
OJ
-M
r^
p
.
0
CTl LO f=3~ CM LO J"-^
in CM ro oo CM -o
'-O *^J- JcM CO «^" =3"
""> CM 'Xl CM
<*
'Xj co f^ -- co jro
o co } cvt co r°
ft. , i .
r-^. ' *Lo >^ i r\
O) CTl
r- L_ ^
e => o
LO cu
O S- O)
> + ^
t
LO r^- JCM LO -=J }
O LO f* OO LO ^>
co ' JLO i ' cn
^3" CM JCM <* CM |
to ^
i
a>
c
i
n3
-4_J
C
O
( j
cu
cu
o *
N)
CU O
^ 1
ro JQ CM
O CU "3 i CL)
13 -t-3 r « ^3 -4-J
§C >r~ r c >r"
ai-MLj- cu cu -4-> >+_
C>tnO U-i >V)O
r- ai o J- cuj->cuos_
E i- U Q. V--Ml_OCL
»~ O O
LO CM K^
'-O LO jCTi
CM LO ro
CM CM 1 1
*S*
LD t_O 1
«3~ CM ICVI
ro cu
i r
O | i
fl3 4_> CU
r O 3 +->
r jQ C ''
r- cu -*--> l+
^ i^ > to o
cu c cu o s_
s_ T- s- u a.
C l_
o -o
z.
30
-------
labeled by the beverage producer, labeling cost could probably be signifi-
cantly reduced. The total handling costs and revenues by activity are shown
in tabl e 6.
5.4.4 Net Profits
The above estimated profit rates ar£based on the inclusion of the
major handling costs only. In addition, Curry Co. indicates the existence
of other costs for both equipment and supplies as well as supervisory and
management time. Sane of these costs can be accurately identified, however,
most must be estimated. TaDlt 7 summarizes the economics of the deposit
experiment using both the actual return rate '0.69) and alternative return
; at es.
Based on total beverage sales of 54,325 cases and the assumption that
98 percent of sales were in aluminum cans with the remainder equally divided
between nonref il 1 abl e beer and soft drink bottles, a gross profit of $14,304
was made during the first summer.* However, Curry Co. has estimated that
about $6,298 was required for equipment and supplies as well as for super-
vision and management of the deposit experiment. The labor time ($4,191) is
an estimate of the cost of the time spent by Curry Co. supervisory and man-
agement personnel. This time was supplied by the regular Curry Co. em-
ployees, so no actual out-of-pocket expenses were incurred. However, since
they were devoting time to overseeing the experiment, some other activities
were getting less attention than they otherwise would. The result of sub-
tracting these costs results in an expected profit of $8,006 for future
summers assuming the costs, productivity, return rates, and sales volume
are the same.
*This value differs si ightly fran the value shown in table 6
$14,160) primarily due to rounding errors.
31
-------
Table 6. Total handling costs and revenues by activity
Activ ity
Handling costs
Label 2,113
Collect deposit 483
Return deposit 2,714
Dispose of containers 6,170
Accounting 277
Subtotal $ 11,757
Revenues
Forfeited deposits 20,209
Scrap sales
Aluminum cans 5,554
Steel cans (bimetal) 87
Nonrefillable beer bottle 67
Subtotal $ 25,917
Gross profits $ 14,160
32
-------
c»
-t-J
("
>^
;/,
F;
Q.
0,
Q
1)
w-j
;
ir
^'
i
M
o
^
C )
r~
£-
b
c:
0
u
CD
.c
-t >
i
"o
>-,
S-
E
E
13
OO
rJ
OJ
,
-Q
«\
r- \ CXI CXI
1 I I 1
t
CD CD cxi |CXI r^ O CD to LO O
«^f LD Cxi to O O < f^~ ^J-
CXI -*~> CXlCXJCXjCDi ^j-
| 1 1 1 ~ r. »
' CXI CXI
LO
r^ cxi
LO LO
*» **
ro to
r~
1
<*
O LO
< cxi
tO LO
«* «*
to to
1
Lfji
C7^
O
r.
O
CXJ
i
LO
CO
"
r-^
1
fe^> fa^ fe*»
LO O j'-d r-- O O to LO O
f~-- cxi Ln kj- to O CD i r^ «d"
LD jto cxi cxi cxi O ' *\ r.
UD |kO ' CXI CXI
00 LO
=d- cxj
CO LO
«*
to
r^.
r *
f~"
«
10
<* ^ ^
^OOCvlfcO r^-O OtOLOO
cOOcxio too Or-r--«d-
00 ^- r JD CXJCXJ CXJ O r ^]-
«\ «. r> ^ ^
O LO
r- CXJ
O") LT5
*- «i
ro ro i cxj r-^ , i ~- to
r~ \ i
LO
oo
i ^
<^)
co
o
10 *=^- ro ro ^ O O to LO O
crvr^CTiUD too Oi r^-«d~
i . < -O CXICXI CXIO' ^f
o CD i cxi cxi
CXI |CXI
Li-) LO
to cxj
CXJ LO
W to^
1
o
^j-
r--
r-^
to «<» v^
o
Ln
CD
*
CXJ CTt LO -O r . CD CD to LO CD
i
t?*> f> +f}
* -V
OJ CTi
-^ C ^
'i ^ d
QJ 4-> ^-- -
00 QJ
O S- QJ
13
*^_
-K
cor-^crv==j- r~~o OtOLoO
toocxio too Oi r-^«3-
o . i no cxjcxi cxj o ^t-
n ^ n *i «*
=3- =^ i cxi cxj
tO LO
O CXI
0^ LO^
oo to
i
oo
sj-
^
4y£$ &^ ^^
,U
^3
O t|~
-Q O 0>_
CD oo S- O
r ^XL 4-J OJ i_
+-> C 0000 I T30JOO 1~
*->- OJO -r- OJECU O
O S- 4-^ O "^ (& ) ^ c | * Q c tj
f""i TU (O C &_ O 5 f3 -4-3 TO
E OVr *O 4-) CD OO £ S_ 4-* U
S- 4-> -r- C QJ Q. -i Q) CD-r- O
QJ tf 4-* 4i "O 4-3 QJ XI- 4-* H3 O 4
(DO OO S_ C C O> QJ QJ CO "O S^ O i_
-Qoo QJ i_ rtJ -i- 03 Q.QJ CXl_OJ
3S-001- 4-> OOQ-E
OJOJ -OooD.C7)O QJOJ 4->O) E
C i OQJQ. (134-> C4J OoO-C4J3
4->OfOn3 ^£ O4-)T3i_ Oc
(^p, ^Z j_, (^ t_>o'lioQ^ i_>4->O'Ol/)
O 3 T- T- r S_ , 00 QJ T" E £ QJ ' S- C S-
i_ C o^. i+_ (T3 13 *O QJ i- O r > QJ jZ. > *[ ^3 QJ f*i (/^ »r~
Q- ! QJ QJ 4_i f 3 Z3 m QJ O *O t_ O^ O*OO c } E 3 QJ *i
E S- S- O C d 4-* S» 4^ QJ fO M O i_ QJ 1 >
OO 13 C C 4-> 00 C O OO C7> C O. C '*" *"" Q- OO CTi 4-J TJ Q^
Ot3c:c: c OOQ_ i-ooE3 c c r (TJ ( ^
ts s: o z: E oo
-
"CD
OJ
XJ
13
r-
c:
r
"£3
c:
01
1
03
4-)
OJ
E
i
_o
OJ
tj
"*)
tj
S.-
C3
4-J
O
^f-
un
f»
CD
£:
i
4-J
^3
O
cx
1
i-
4.3
r
03
c:
c>
r-
-4-J
i
T3
c:
^C
K
33
-------
In addition, management time was also expended due to start-up and
the experimental nature of the system. Subtraction of the estimated
value of this time ($6,525) leaves a net profit of $1,481. There were a
few other costs which are not included here for phone calls, trips to
EPA in San Francisco by Curry Co. personnel to discuss the program
implementation and the cost of a shipment of bimetal scrap. If all
these additional items are also included, the deposit system for the
first summer would break even.
34
-------
BIBLIOGRAPHY
Alpha Beta Acme Markets, Inc., "Bottle Survey '71." La Habra,
California: 1971.
U.S. Department of the Interior, National Park Service, Campground Use
in the National Park Service; Camper Days in Areas Administered try
the National Park Service; Public Use of the National Parks,
(monthly) , and Public Use of the National Park System, (semi-
annually). Washington, D.C.: U.S. Government Printing Office,
1974.
U.S. Department of Interior, National Park Service, Yosemite National
Park Collection and Analysis of Qualitative TransportatTon Data7
Washington, D.C.: U.S. Government Printing Office, November 1973 .
Yosemite Park and Curry Company, 1975 Annual Lnvironmental Report.
Yosemite Park, California:
Youngblood, C. E., and T. H. Bingham, A Guide to Implementing a Deposit
System. Reseach Triangle Park, N .C . : Research Tri angl e Insti tute,
1977 (draft report).
35
-------
00 ^-~
C 00
oo on
O) »->
~^T C*
, y . . . . .
f J" -C T J" 7 X
C -t X X o"i
7- ~ ."V *"! "-- -O
V-
oo
C
o
in -K
O)
S-
ra
4-"
c
on
QJ
OJ
OJ C
OJ
01 OJ j
t- j
CD
QJ
-C '
>i .
ra !
0 I
O
ry; "c -X rvj -v, -y »X "V
o o o o o c-
o o
co
ocr
A-2
r> r; o = o o
o o o o
-------
(/) .-
c
S_ 4J
t/> l/l
»->
c o e o
"trc.-'rcr^oO'OO. scr~r j>
;r O c; o c 3 ^ »oc^'^
~ r X .C- ~
-c ~
1*-
i .1 ~~ ^r -J c
-x .^
en
0
cu
i_
-a
c:
v>
C
O) i
OJ C
C a. w
!! ^ ^
'! ^fe
:i ^
"vi r
/- ~
x ^
' ~- -r r ."
- -X r -f -~
J ~sj V ">. X-
"V. 1 ~
^ .". f-
X1 -.1
>,
o
C
. Aj "^ T J^
c x »"i .T /> .r
j-crAj^nrfL-^-T^ £
fV "v. X .-vj X -XJ c
-?OOO^C?O DOO
2
-O -C
4? C
o ~ ">roooc
o r> s r>ooo uooooo
A-2
-------
co .-
C co
S- -M
3 -r-
M C
O) 3
a: -
c tr ~ ~> o
-3 c
cc -C X CT-
COO >OOO OOOO OOOOOOO O O
o o o o o o o o o o o o o o o o o o
*\i o \J»»»»«»r--'»»»»
ri'.P'XiX-OJ"'^-*'". "i^JfVrtf"^''! 30 LT> -O *-
x x _'^
£__ < * x x
0- «
-- » oo cc
o- r-
(LI
1/1 I
t/i
o LI j^
c.- o r\j o ir*
s_
O
01 «
(T
O-
(O
4->
o
OJ
s_
D
- .T ~U .O ' f\J ? c
"VJ "O
"U TV,
-V
-VJ
l/l
QJ
O)
JC
(II
4-> S-
cn
"5 O t5 C O ^C-~"DOO
J^ J~' -P O O O O O al J"> -P
oooocr
oooooo
- J^ J"l
r>ooooooooo
J^ lT .0 J~ J^ O O O O J1
"Vj C
X
~\j o ^r - x CT- o
OJ
(T3
(t3
Q
C O O O C r
O O ^ .3OO
oocooo
C O OC. D
OOO3O
-3 300000
OOOOOOOOOOOOO
S3OOOOOOOOOOOO
"\J
o
oo o ooooooooooo^oc>ooooooooooooo
OOOOOOOOOOOOOOOOOOOOO'OOOOOOOO
O OO OOOO OOOOOOO O OOOOOOOOOOOOOO
oc-oooooooooooooocoooooooooooo
OSOO30OCOOOOOOOO--OOOOOOOOOOOOO
A-3
-------
»-> C
^) 3
c o o o r- c o
Or? o O O 3:
o c r cr
"^7 "3* ^ ~P C* J"^ "^
~ o --- c j- r x
./* ;r o -3 c; 'v, a-
x: j~< < -^ £ _n
~^. r* r\i f *»- .
^. W VJ -, *. *^
-r -: c-
c ~- r
^ LT J"' ^ J~>
(O C
AJ ^
f - x
~ ^ r>
^- > 7- -x z.
c a- >- x
C (7- - J~ i
cr x -- ^ ~
c
o
C_3
CTl
c -. - r .- : r - .- - c ^ -.
OJ
s_
TD
O)
r- ''
a ''
+j
o
O)
x:
OJ C
S- OJ
a> r: i-
cn-M x:
fl3 H3 *T3
i i- i*
ai ai
> a. to
< E a)
oj ai
en
a>
V ^ 'J "O 'Xl »J ~s* V ^w X ~V O VJ V '^ ~V, "w X.' ^_ Vj '0 'X, *J ^' VJ
I J-
J~' ~ ~
: L.'>~^^-r'--'t
-- .. -r c r ^ _ ^ r. i- x
c
O
o r>
' t 7~ o « » ;r «* '"Y; -^ ~ iTi ^> ~^ .{"."* -^ X o Tt t/> j
1 \j "Vj *<1 <"! Aj-^TTJ^C'^i^,! .. ^'\J"VJ'AJ'Xl'Vj"Vjn
^> O1 O "^> .^) O O O O O T O O O .O O O O d O O -O
^ O O O O O O O O O O O O O O O ^ C5 O O ^ O
tr o o> o o o o o o o o o o o o c cr o o o o o
;>OOC.DOOO3OOCC OOOC'C'OC-O:?
ooooo o'oocooc-ooocDOO-crooo
I
; r
-C X X' X
C X' X X X- X' X.
o o o o o o c
o o o o o o c
c; o o ,D
-------
1
1 ! " IT O -Z^ 3- O CT « O O O O O" O O" O O CU. G C2 O
! t/1 - i o o o. o o »ocoooooooooooo
4j cr ' r ^^c>^^c>---'>'^o £'"?-£. £''\>tTw'X"3'CT*otr'
OJZ3 '; "" "^ IT C T ^ ~ ~- C -X' ~7 -^ O" ' _T '''I O" 3" J"> O sO
C£ * ' i I - C^ X X *""- ' T~ X -/" "^ r?"\jxC'Xl"XJ'v"*'XJ::3'XJ'sri
11 1
'i 1
1 1
1 t
t
ll '
1
1 i
_.-,'- - -- - - <~- _-.<-,- O
, * "~ | »*»*" c: > -I, ~ o o o o c *
,0 c- i /- --, ------ -. r: c" 'X ^ c * -"- :r ^ *-* c? « -^ j"» oj
j i/> 3 j ~r- - -\ 'v -, c -* - r x ^ *- t > .-T ^r O *^ "u o
, . r ^ r ^ -" xj -^ -n 'xi -^ .n
; 1 !
!
^ i J
1! ^ \
3 ' 0
V ( S' j
1 ;i -r~ , ~ "^ -x "v. -^ r ; c "-" "«. . : .* -r ' T c 3"
>>) i -^»--.c3j-ccv-^'----^-rj;--^-.r-^o.O'-occ
_ r \_j i '"' ~* v i "^ ^ V T" "V. "^ "*'"' *. - ^X.: "*" ./> IT *s*^ ^7 s v>
_} Q I VJ S. V V>
L) H- 1
t
^ li z? i
-o OJ I
U _C '
i cu c ;
yi J V-'
CD i
QJ \
^ "O
^ >. -
O)
-° 1 i
H- ! _ , ^ ^ - ^, .. _^ . , -. _ _
1 " "- "^ "Z> ~ O O O T> ^ 3^ ^. ^» ^ O O O O O "!> O
1 ~'-"v j~^-^'CT"DOC^^'^1^CJ7* O O O O O ID ^
>^ * ., --> .,. -,, . ,^_~. ^^ -^ _. ^
^3 ' J "*^X ,~^"I.^_^ * _-^ w^
' "^ *^ ^
t
<
i
O CT CT O^ "D O O O O O. O O O O O O O O O O O ooooo
II *-" oo^Cioooooooooooooooooc-
£ r-^o^ooooooooooooooocoo
-g2 r^OOOOOTOCOOOOOOOOOOOO'
[ ^- ^ _,- ^. ..^^ -. ,» ,^^ ,-»^ .^^ .^.^ j^ ^^ .^^ ,»^
i
o>
c~
*->
t~
to
cu
r
0
JC
cu
^^
i_
cu
o.
'>
LO
CO
c
o
"O
O)
ii
^
5-
its
Q.
^^
^-^ 3^>
i-
O>4-»
c c
r- CU
*~
4-* O
c
-J
CU XJ
S-
4" O
O +J
t- (/>
O) <-
^ > CO CO CO r
E ....
c 01 *a- «tt co co
Q.
O)
c~ ^0
4J i- O)
3 VJ
E O C 4J
i- 'u i. «/>
«4_ n
4«> CU ! c** fi IO
O »-> Ol
E "3 V- 3 O> O
4-> O>O 00 OQ H-
1/1 C
*>
o
CO
^
"t")
1
0)
r
ex
0
o>
a.
4*
cu
CT)
o
cu
o
(0
"O
Ol
c
to
4->
jQ
o
r~~
t)
j *
o
>,
^3
T3
1
J_
o
CO
»
>
o
J >
"8
o
TD
(O
cu
tO
Q
in
'~~
^_
o
in
o>
Ol
*>
o
a.
^
Ol
>,
i.
J-
3
o
a
c
-------
HI <-
S_ Oi
"3 OJ
S- 5~
o; _c
O TC>
E u_
O)
4-i i/i
OJ
OOOOOOOOOOO 3OOOOOOOO
C: O O O O O 3OOOOOOOOOOOOO
o *- cr x -C c:
C -C
£ ij"1
C. -O
OJ cn
5: a;
T3
OCJOC-OOGOOOOGOOOOOOOO
OOOOOC^OOOCC'OOOOC'OOOC
OOOGOOCOCO
i-^-v-i-~^ C -C C-C1^-1^
r- o "vi « 'u ra JC^O^T
XI ^J -O -O "Vi "VJ ? ^ C 3
cooooooocro
'^i.r>-Oox'\jLn-ocC'cr
,O _r- --5
C
I/)
c
cu
c
o
"X
\i
"VJ
."VJ
CO
OJ
-O
o ooooo oocsoo
ocooooooooooooooodoo
OOOOOOOOOOC'OOOOOCOOO
oc oooooooooooooooooo
oooooooooooooooooooo.
oooooooooooooooooooo
B-2
------- |