United States
Environmental Protection
Agency
Risk Reduction
Engineering Laboratory
Cincinnati OH 45268
Research and Development
EPA/600/S2-89/032 Feb. 1990
James K. Edzwald and James P, Maliey, Jr.
The goal of this research was 'to
investigate dissolved air flotation
(DAF) as a drinking wafer treatment
method. Several different water types
were examined; low turbidity waters,
waters containing aquatic humio
substances, and waters with algae.
The study design Included synthetic
waters so that raw water quality char-
acteristics could be controlled and
natural waters collected from three
water supplies. Flotation studies were
conducted using a bench-seal® sys-
tem of four parallel units.
The study showed that flotation
performance depends on raw water
quality, prelreatment, bubble size,
and bubble volume concentration
and'that DAF is "effective in treating
water supplies containing humic sub-
stances or natural color arid supplies
with algae. To remove particles, the
particles must first be destabilized
with coagulants. A short flocculatlon
period ahead of flotation may be
beneficial; however, long flocculalion
times are not necessary. That DAF is
an efficient process is due, in part, to
the small bubble size produced by
the process (bubble diameters of 10
to 100 pm). The bubble volume con-
centration, a design and operational
parameter, is controlled by the sstu-
rator pressure and recycle ratio.
Typica! values of 70 pslg and 8%
recycle produced bubble volume
concentrations in this research of
4600 ppm. Residual turbidities were
significantly lower for DAF than for
conventional treatment for all raw
water systems, particularly at colder
water temperatures. For water syp-
piies containing humic substances,
there was no significant difference
between DAF and conventional
treatment In the removals of U¥
absorbance, true color, DOC, and
dissolved organic halkte precursors
(THM and TOX). The removal of
dissolved organic matter depends on
coagulant chemistry (i.e., removing
the dissolved material from solution
Into floe particles) and not on the
solid/liquid separation process.
This Project Summary was devel-
oped by EPA's Risk Reduction Engi-
neering Laboratory, Cincinnati, OH, to
announce key findings of the research
project that is fully documented in a
separate report of the same tills (see
Project Report ordering Information at
back).
Introduction •
In Europe, DAF is an alternative' clari-
fication process to sedimentation in
drinking water treatment. In the United
States, DAF is not widely used although a
few communities have recently installed
DAF plants. This project was needed for
several reasons. First, there has been
little research on DAF; consequently, the
design and operation of DAF plants are
based largely on prior experience and
pilot plant studies. Second, the 1986
amendments to the Safe Drinking Water
Act require EPA to establish criteria for
filtration of surface water supplies. EPA
has proposed a filtered water turbidity of
0.5 IMTU or less. Pretreatment ahead of
the filters is an important step in potable
water treatment, and DAF is an alter-
native pretreatment solid/liquid separation
step to sedimentation. DAF should be
-------
more efficient than sedimentation in
treating waters with low density particles
such as algae, alum Hoc, and floe.
produced from treating low turbidity,
colored waters. Third, the high hydraulic
loadings (5 to 15 m3/m2/hr) and short
hydraulic detention times (e.g., 10 min) of
DAF tanks compared with, conventional
settling tanks can provide considerable
capita! cost savings. Finally, the ability of
DAF to' remove natural color and
trihalomethane (THM) precursors needs,
investigation.
The goal of the research was to
investigate DAF as a method of treating
two types of water supplies: low turbidity
waters containing natural color and water
supplies with seasonal algal problems.
Specific objectives of the study were:
® to compare DAF with conventional
water treatment of coagulation-fioe-
culation followed by gravity settling
(CGS)for particulate removal (i.e., tur-
bidity as well as particle counts for
supplies with algae);
• to -compare DAF with conventional
water treatment for removals of color,
dissolved organic carbon (DOC), tri-
halomethane (THM) precursors, and
total organic halide (TOX) precursors;
• to examine the effects on DAF per-
formance of raw water quality, water
temperature, chemical variables (pH,
coagulant type,' and coagulant dose),
flocculation period before DAF, and
DAF design and operating variables
(flotation time, saturator pressure, and
recycle ratio);
e to evaluate alum and polyaluminum
chloride (PACI) as pretreatment coag-
ulants for DAF. •
Procedyre
Several types of synthetic waters and
natural waters (actual drinking water
supplies) were used. Synthetic waters
were used so that raw water quality
characteristics of low turbidity waters and
waters with aquatic humic substances
could be studied at desired concen-
trations. The synthetic water systems
were: waters prepared with aquatic fulvic
acid at concentrations of 2 to 10 mg/L as
DOC (true color values of 20 to 100 Pt-
Co color units); water prepared with
montmorilfonite clay (clay concentration
of 20 mg/L and turbidity of J .3 NTU)
without fulvic acid to simulate a low
turbidity water without color; and waters
containing low turbidity from the mont-
morillonite clay and fulvic acid (at con-
centrations of 2 and 10 mg/L DOC).
These two fulvic, acid/clay waters were
used to test, low turbidity .water with low
and high concentrations of aquatic humic
substances. The fulvic acid was extracted
from a stream (Provencial Brook), which
-"feeds Bickford Reservoir, a water supply
for Fitchburg, MA.
Three natural waters were used.
Washington Mt. Brook was used as a
supply of high natural color. Washington
Mt. Brook, which is a future water-supply
for the towns of .Lee and Lenox, MA, has
low turbidity (1 to 5 NTU), low alkalinity
(3 to 20 mg/L as CaC03) high true color
(50 to 200 Pt-Co color units), high DOC
(5 to 12 mg/L), and high THM and TOX
precursors (7-day, TTHMFPs and
TOXFPs of 500 to 1,600 pg/L and 880 to
3,600 jig/L, respectively).
Upper Root Reservoir was used as a
low turbidity supply with little natural'
color. It is an existing water supply
source for Lenox, MA. The raw water tur-
bidity is typically 1 to 2 NTU, whereas
the true color is generally less than 20 Pt-
Co color units.. it has low concentrations
of DOC (1.3 to 2.2 mg/L), TTHMFP (100
to 200 ng/L), and TOXFP (200 to 350
iig/L).
Wachusett Reservoir was used as a
supply for studying the use of DAF to
remove algae. The Wachusett Reservoir
water is a supply for Boston and sur-
rounding communities. It is a high quality
supply (Boston currently does not filter)
of low turbidity {0.4 to 1.1 NTU), low DOC
(2 to 3 mg/L), and low THM precursors
(TTHMFPs of 130 to 160 pg/L). Two
different algae, Chlorella vulgaris (green
alga) and Cyclotella sp. (diatom), were
spiked into Wachusett Reservoir water
under controlled growth phase conditions
to study the effects on DAF performance
of algal growth phase and cell number
concentrations.
Two coagulants were investigated,
alum and pofyaluminum chloride (PACI).
The PACI, prepared in our laboratory,
contained approximately 10% monomeric
aluminum and 90% polymeric aluminum
species. The dominant fraction of the
polymeric aluminum was probably in a
form such as AI13O4 (OH)247*- Jar tests
were performed on all test waters before
the DAF studies to determine pH and
coagulant dosages required for removal
of humic substances from solution, and to
determine coagulation conditions such
that the floe particles were unstable (i.e.,
coagulant dosages needed for particle
destabilization),
Flotation studies were carried out usiruj
a bench-scale system of four parallel
units. The synthetic water studies
involving fulvic acid and montmorillonite
clay were conducted in two stages. In
Stage I, a general statistical factorial ex-
perimental design in which two levels of
each of the variables were studied simul-
taneously: pH, temperature, DOC con-
centration, clay concentration, floccu-
lation time, DAF detention time, and
mode of separation (i.-e., DAF vs. gravity
settling in conventional treatment). Stage
II studies were designed to investigate in
greater detail the effects of coagulant
dose, flocculation period, temperature,
and bubble volume concentration (con-
trolled by saturator pressure and recycle
ratio) on DAF performance. The per-
formance variables measured included
residual turbidity, residual dissolved
aluminum, and various measurements of
dissolved organic matter including UV
absorbance (254 nm), DOC, TTHMFP,
and TOXFP.
The DAF studies of Upper Root
Reservoir and Washington Mt. Brook
examined coagulant dose, pH, floccu-
lation period, and season, (raw water
quality and water temperature). Flotation
time, recycle, and saturator pressu
were normally set at 10 min, 8%, and ',._.
pslg, respectively.
The DAF studies involving spiking of
algae into Wachusett Reservoir were de-
signed to; (1) define coagulation condi-
tions for DAF treatment; (2) evaluate DAF
performance for removals of particulates
(algal cells, turbidity), dissolved organic
matter, and THM precursors; and (3)
examine the influence of algal growth
phase, algal type, algal cell number con-
centration, flocculation period, water
temperature, and bubble volume concen-
tration on DAF performance. Algal cul-
tures grown in the laboratory under batch
growth conditions were spiked at desired
initial cell concentrations and growth
phase into water collected from the
Wachusett Reservoir. Chlorella was used
under both log and stationary growth
phase conditions to evaluate the effect of
algal growth phase, and Cycfcrfe/fa was
used under declining growth phase
conditions. Standard cell concentrations
of Chlorella and Cyclotella were 10s and
5x104 cells/mL, respectively. A floccula-
tion period of 5 min was used in standard
experiments, followed by 10 min of
flotation using a bubble volume concen-
tration . (4>b) of 4,600 ppm (i.e., saturate
pressure 70 psig, recycle ratio 8%). (
-------
This study established that coagulation
is needed for effective flotation as
illustrated by the results in Figure 1.
Figure 1 shows DAF performance for
treatment of a water containing aquatic
fulvic acid at a concentration of 5 mg/L as
DOC. The results are presented as a
function of the coagulant dose, PACI in
this case. The UV and DOC data show
poor DAF treatment without the addition
of coagulant and for low PACI doses. The
optimum PACI dose occurs at approxi-
mately 2.5 mg/L, Higher PACI doses
cause an increase in -turbidity indicating
overdosing. The electrophoretic mobility
(EPM) data indicate that the electrical
charge on the floe particles is near zero
at the optimum dose, which in turn,'
indicates particle destabilization. Sepa-
rately run jar tests indicated the optimum
PACI dose .was 2,5 mg/L, which is
identical to the DAF results.
For all waters tested, the optimum
coagulation dosage for particle destabil-
ization based on jar tests was found to be
the same dose required for optimum
treatment in DAF. This means that jar
tests can be used to determine coagulant
dosages for DAF, thus minimizing the
number of tests in pilot plant studies, and
can be used to establish coagulant
dosages for full-scale plants. The results
also indicate that coagulation removes
aquatic humic substances from solution
by a phase change, from soluble organic
matter to floe particles. The particles are
formed by precipitation of aluminum-
humates or by precipitation of aluminum
hydroxide with loss of humic matter from
solution by adsorption on the aluminum
hydroxide. The mechanism of particle
formation depends on the coagulant and
the pH conditions. .
The study showed that the flocculation
period had little effect on DAF per-
c
"5
i
a
s>
TOO
80
60
40
20
0
. .1-5
j? 1.0
0,5
0
2.0
^ 1.0
o
« 0
i
"~ -1.0
-2.0
—. a—— — ~ —-a
I
1.0 2.0 3.0 4.0
" PACi Dose (mg/L as Al) •
5.0
"•gure 1. Effect of PACI dose on DAF for 5 mgIL fulvic acid, as DOC (pH 5.5, 25°C, 20-min
j flocculation period).
formance. Figure 2 illustrates this for two
water temperatures when PACI was used
as a coagulant to treat water containing
fulvic acid, DAF studies with alum
showed similar results; however, residual
turbidities were generally higher at 4°C,
which indicates water temperature has an
effect when aium is used as a coagulant.
Further analysis of turbidity data
indicated that PACi produced lower
turbidities than did alum for either DAF or
CQS (conventional treatment with settling)
treatment under cold water conditions.
This is shown by the data in Table 1 for a
variety of synthetic waters and for
Washington Mt. Brook. The only
exception is the one experiment for
Upper Root Reservoir,
The effect of bubble volume concen-
tration on DAF performance was exam-
ined. Bubble volume concentration was
varied by holding the saturator pressure
constant at 70 psig and varying the
recycle ratio at 2%, 5%, 8%, and 12% to
yield bubble volume concentrations of
approximately 1,200, 2,900, 4,600, and
6,600 pprn, respectively. For all of the
waters tested without montmoriilonite
clay, a bubble volume concentration of
2,900 ppm or less was adequate for good
DAF treatment. For waters containing the
clay at 20 mg/L, a bubble volume con-
centration of 4,600 ppm was needed for
effective flotation which indicates more air
bubbles are needed to float these high
.density particles.
This study found that the removal of
soluble organic matter (measured as
DOC, TTHMFP. and TOXFP or by a
surrogate measure such as UV absorb-
ance) is independent of the solid/liquid
separation step, i.e., DAF vs, conventional
treatment with settling. This is illustrated
by DOC data in Figure 3 for Washington
Mt. Brook, The figure contains a bivariate
plot of the data. The 45-degree line
represents the "line of equal value"; data
points falling on this line indicate an equal
response for each level of the
independent variable. Each data point
represents the results of paired
experiments for identical pretreatment
conditions followed by DAF or sedi-
mentation. That the data fall on the 45-
degree line indicates no difference
between DAF and conventional treatment,
This is an important practical result. DAF
treatment has no advantage over con-
ventional treatment in removing non-
volatile soluble organic matter because
removal from solution is due to
coagulation. Therefore, for the same
coagulant dose, DAF and conventional
treatment give the same removals of
-------
2.0 --
1.S--
1.0-
0.5--
-p
0.0'
2.0 ~~
_ b-
f.5--
10--
0.5--
0.0-
0.5-
O Before DAF
Q /Wer DAF
»&••
25 "C
-0-
...-£!
O Before DAF
a After OAF
4 °C
-B-
...Q..
I
0.0>
-0.5'
a 4 °c
O 25 °C
--•B....,
JO 15
Flocculation Time (mm)
20
Flgws 2. Effecf of temperature and flocculation time on DAF for 5 mgtt fulvic acid, as DOC (pH
5.5, PACI dose 0.5 mg Alimg DOC.
dissolved organic matter. The two
processes do differ, however, in their
ability to remove the floe particles pro-
duced by coagulation (see Comparison
section below.
Experiments examining the effect of
coagulant dose on DAF performance
verified the results described earlier: little
or no removal of turbidity and algal cells
occurs without coagulant addition, and
optimum removals by DAF occurs at the
optimum coagulant dose based on
particle destabilization. Figure 4 illus-
trates this for alum for DAF treatment of
Chlorella wulgaris at two growth phase
conditions (log growth and stationary
growth). The data also show that the
optimum alum dosage is lower under
stationary growth phase conditions than
under tog growth conditions as well as
better removal of algal cells. Similar
results were obtained when PACI was
used as the coagulant.
DAF removal of Chlorella cells was
studied at two water temperatures, 4°
and 21 °C. Flocculation periods were at 0,
5, and 20 min. Alum was the coagulant.
At 21 °C, ftocculation period had little
effect on DAF performance. At 4°C, a 5-
min ftocculation period improved the
removal of algal cells and turbidity when
compared with no floeculation period.
There was little benefit, however, in
increasing the flocculation period. The
results indicate that good separation of
algae by DAF occurs .without aggregation
of cells to large floe sizes.
The effect of bubble volume concen-
tration on DAF treatment for three initial
CMorelta cell concentrations is, given ir'"
Figure 5. Good results were obtained witK
the lowest bubble volume concentration
(1,200 ppm) for Chlorella at concentra-
tions of 2x10" and 1Q5 ceils/mL At 5x1Q5
cells/mL, a bubble volume concentration
of 2,900 ppm was needed. For water
treatment applications, algal cell concen-
trations would typically be below 10s
cells/mL so that a bubble volume con-
centration of 2,900 ppm (5% recycle, 70
psig saturator pressure) would be more
than adequate assuming similar alum
dosages (10 to 15 mg/L).
Of DAF With
DAF was compared, side-by-side, with
CGS under the same experimental con-
ditions. Experiments were done for syn-
thetic and natural waters. In ail experi-
ments, the optimum coagulant dosage
was used. Rapid mixing was for 2 min (G
of 380 sec-1), and flocculation mixing was
for 20 min (G of 10 sec-1). For flotation, a
saturator pressure of 70 psig and recycle
ratio of 8% (bubble volume concentration
of 4,600 ppm) were used. The flotation
detention time was 10 min, whereas the
CGS settling period was 60 min. The ratio
of flotation overflow rate to settling
overflow rate was 6 to 1 except in thosi
experiments with the algae water systems
where the ratio was 10 to 1. These ratios
are typical of practice.
In Figure 6, the comparison of DAF with
CGS is illustrated for waters containing
atgae. The data (y-axis, DAF data; x-axis
CGS data) are below the "line of equal
value"; this indicates better DAF than
CGS performance. DAF removals were
99% to 99.9% (2 to 3 log reduction);
conventional treatment was 90% to 99%.
In general, for all the water types
studied, DAF produced significantly lower
turbidities than did conventional treat-
ment, particularly under cold water condi-
tions. .When PACI was used as a
coagulant in either DAF or CGS
treatment, the treated waters had lower
turbidities under cold water conditions
than when aium was used as a coagulant
(see Table 1).
This study showed that flotation
performance depends on raw water
quality, pretreatment (coagulation and
flocculation). especially chemical pre-
treatment, bubble size, and bubbi.
-------
1. Comparison of Residual Turbidities for Mum and PACI at Cold Water Temperatures
Turbidity (NTU)
Sample
Synthetic Waters:
2 mgIL DOC FA
10 mgil DOC FA
2mg/LDOCFAI
20.6 mg/L clay
lOmg/LDOCFAl
20.6 mg/L clay
pH
5.5
7.0
5.5
7.0
5.5
7.0
5.5
7.0
Temp
ro
4
4 to 5
4
4
4
4
4 to 5
4
Alum
3.41
0.70
3.24
8.80
1.00
0.58
4.00
8.41
It
PACI
0.16
0.19
0.94
0.64
1.30
0.45
i.oa
0.80
VV"
Alum
3.82
1.14
3.90
T5.7
1.10
1.19
4.75
8.63
riu*
PACI
O.S8
0.60
1.20
0.51
1.40
1.24
1.52
0.91
Natural Waters: •
Washington Mt. Brook
10/27198
1/14/87
Upper Root Reservoir
2/24/87
5.5
6,0 to 6.3
7.7
7
4
3.20
3.38
1.04
1.60
1.07
1.32
5.50
4.63
1.12
2.20
1.44
1.48
100
00
"o
I
s
o:
60
40
20
|— OAF
Line of
Equal Value
Conventional •
I I
20
40 60
% Removed
80
100
;§ure 3. Comparison of DAF with CGS for percent DOC removed for Washington Mt. Brook
volume concentration. DAF is effective in
treating water supplies containing humic
substances or natural color (DOC con-
centrations as high as 12 mg/L with true
color values of nearly 200 Pt-Co color
units were tested) and supplies with algae
{celt concentrations as high as 5x105
cells/ml were tested). To remove
particles, the particles must first be
destabilized with coagulants. For water
supplies containing humic substances,
the dissolved organic matter must under-
go a phase change (soluble matter to
particulates) which is accomplished by
coagulation. A short flocculation period
ahead of flotation may be beneficial;
however, long flocculation times are not
necessary. Pin-point size floe of 10 to 50
jim should be introduced into the DAF
unit. DAF is an efficient process due, in
part, to the small bubble size produced
by the process (bubble diameters of 10. to
100 ym). The bubble volume concen-
tration is a design and operational
parameter controlled by the saturator
pressure and recycle ratio. Typical values
of 70 psig and 8% recycle produced
bubble volume concentrations in this
research of 4,600 ppm. This bubble
concentration is much higher than the
particle volume concentrations for the
waters tested, and ensured good collision
opportunities between particles and
bubbles and sufficient bubbles to lower
the density of the particles for flotation.
Some specific conclusions of this study
are:
• The optimum coagulation conditions for
particle destabilization and particle-
particle attachment (i.e., conventional
jar test results) are the same as those
required for particle destabilization and
particle-bubble attachment (DAF re-
sults). A practical result is that jar tests
can be used to select optimum
coagulant dosages for DAF.
® Experiments at low water temperature
(4°C) indicated that floccuiatiori period
affected DAF performance when treat-
ing precipitated aluminum hydroxide
particles (alum added to clean water
without other particles or contami-
nants) and particles produced by the
alum coagulation of fulvlc acid at a
concentration of 5 mg/L as DOC. How-
ever, low temperature experiments us-
ing PACI as a coagulant of the fulvic
acid at the same concentration indi-
cated that - DAF performance was
independent of flocculation time within
the temperature range (4° to 25 °C)
studied. The differences in DAF per-
formance observed with alum and
-------
PAGI at 4°C support'the hypothesis
that particle stability and particle-
bubble attachment are a function of
the hydrophilic nature of the particles
as well as their surface charge.
Experiments examining the effects of
bubble volume concentration on DAF
performance indicated that raw water
quality affects the optimum bubble
volume concentration. Experimental
results indicated that good DAF per-
formance was obtained for a bubble
concentration of 4,600 ppm {8%
recycle, 70 psig saturator pressure)
when treating a supply containing a
clay; however, good performance was
obtained at lower bubble concen-
trations when treating waters contain-
ing humic substances or algae and no
clay.
Comparison of DAF with CGS indi-
cated comparable removals of DOC,
UV, true color, and dissolved organic
halide precursors {TTHMFP, TOXFP).
These findings support the premise
that removal of dissolved contami-
nants is independent of the solid/
liquid separation process used.
o log Growth Phase
D Stationary Growth Phase
» Comparable levels of residual dis-x.
solved aluminum (RDAI) were also'
found after DAF and COS.
© DAF produced significantly lower re-
sidual turbidities than did CGS in the
treatment of low turbidity waters con-
taining humic substances, particularly
at colder water temperatures. Also,
PACI produced tower turbidities than
did alum for cold waters.
» Optimum coagulant dosages for alum
and PACI were tower for DAF treat-
ment of CWoreWa under stationary
growth phase conditions than under
log growth phase conditions.
• DAF achieved excellent removals of
algae cells, greater than a two log
reduction in algal counts, for both
Chlorelta and CycloteHa, by using the
optimum coagulant dose without a
flocculation period. At 4°C, a 5-min
flocculation period ahead of flotation
improved the removal of algae and
gave lower residual turbidity and lower
residual dissolved aluminum.hatide
precursors (TTHMFP, TOXFP). These
findings support the premise
The following recommendations are
made for future research. First, studies
should be conducted that examine the
integration of DAF with deep bed filtra- .
tion. These studies should compare DA'
with filtration to conventional treatment -
with filtration under optimum operating
conditions for each. Second, operating
and capital cost data should be devel-
oped for the purpose of producing cost
curves so that design engineers can
compare DAF with conventional water
treatment in predesign studies of water
supplies. Finally, the characteristics of
sludge from DAF should be evaluated
and compared with sludge characteristics
from conventional settling processes.
DAF may produce a higher solids content
sludge, which would yield cost savings in
sludge treatment and disposal when
compared with that of conventional
treatment.
The full report was submitted in ful-
fillment of Cooperative Agreement No.
CR812639 by the University of Massa-
chusetts under the sponsorship of the
U.S. Environmental Protection Agency.
50
Alum Dose (mg/Lj
Figure 4. Effect of alum dose on DAF treatment of Chlorate vuigaris under tog growth and
stationary growth phase conditions (10s ce/te/mL, pH 6.5, ffoccu/afr'on time of 5 mm,
and 4>b of 4,600 ppm).
-------
4,0 -i
1000 2000 3000 4000 5000 6000
Bubble Volume Concentration (ppm)
7000
Figure 5. Effect of bubble volume concentration on OAF treatment of CMoreifa vulgaris (initial
cell concentrations of 2xl04, 10s, and 5xlOs cells/mL; optimum alum dose at pH 6.5).
James K, Edzwald is with the University of Massachusetts, Amherst, MA 01003; and
James P. Maltey, Jr., is presently at the University of New Hampshire, Durham,
NH 03824,
Kim fL Fox is the EPA Project Officer (see below).
The complete report, entitled "Removal of Humic Substances and Algae by
Dissolved Air Flotation," (Order No. PB 89-214 407IAS; Cost $28.95, subject to
change) will be available only from:
National Technical Information Service
5285 Port Royal Road ' ' .
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Risk Reduction Engineering Laboratory
\ U.S. Environmental Protection Agency
>>) Cincinnati, OH 45268
-------
2.0
0.0
0.0
1.0
Turbidity (NTU)
1.5
2.0
Log N/Na (Algal-Size)
Figure S. Comparison of DAF with CGS treatment for (a) turbidity remaining and (b) removal of
algal cells for Chlorella vulgaris and Cydotella sp, at water temperatures of 4" and
21 °C.
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S2-89/032
^•jvj-j^UaQFFICIAL. MAI,"
000131433; ps
fff '89193-3JI78
------- |