ENVIRONMENTAL PROTECTION AGENCY
           \
        OFFICE OF ENFORCEMENT
             EPA-330/2-77-002
DESICCANT USE OBSERVATIONS
     ELLIS COUNTY,  TEXAS
          (SEPTEMBER 21-27, 1976)
     ENFORCEMENT INVESTIGATIONS CENTER
          DENVER, COLORADO
              JANUARY 1977

-------
      Environmental  Protection Agency
           Office  of Enforcement
            EPA-330/2-77-002
        DESICCANT  USE  OBSERVATIONS

            ELLIS  COUNTY,  TEXAS


           September 21-28,  1976
               January 1977
National  Enforcement Investigations Center
             Denver, Colorado

-------
CONTENTS
I. INTRODUCTION. . . . .
.......
. . . .
. . . .
. .. )
11. SUMMARY AND CONCLUS IONS.. . . . . . . . . . . . 4 . . . 3
General Conclusions. . . . . . . . . . . . . . . . . .. 3
Specific Conclusions. . . . . . . . . . . . . . . . . . 3
III. DESCRIPTION OF STUDY AREAS
.........
. . . .
.. 7
IV. USE OBSERVATIONS. . . . . . . . . . . . . . . . . . . ." . 11
Pre-Application. . . . . . . . . . . . . . . . . . . . . 12
Application. . . . . . . . . . . . . . . . . . . . . . . 20
"Post-Application. . . . . . . . . . . . . . . . . .'~ . . 22
V.
EVALUATION OF METHODS. . . . . . . . . . . . . . . . . . 34
On-Site Observations. . . . . . . . . . . . . . . . . . 34
Air Sampling Devices. . . . . . . . . . . . . . . . . . 34
Tracer-Dye Studies. . . . . . . . . . . . . . . . . . . 35
Spray-Droplet Cards and Slides. . . . . . . . . . . . . 35
Environmental Sampling. . . . . . . . . . . . . . . . . 35
REFERENCES. .. . . . .
. . . .
..........
. . . 36
APPENDIX: Sampling Devices and Methods.". .
. . . . . . 37
TABLES
1. Sampling Stations, Ellis County, Tex~s. . . . . ~ . . . 10
2. Arsenic in the Environment, Ellis County, Texai . . . . . 12
3. Formulation and Diluted Material Analyses
.....
. . .18
4. Weather Conditions During Application. . . . . . . . . . 21
5. Droplet Impingements on Thermofax Paper
. . . .
. . . . . 30
iii

-------
FI GURES
" .
Study Area and Station Locations, Ellis County, Texas. . 8
:2. Greenburg-Smith Impinger Unit, Droplet Card Platform,
High-Volume Air Sampler. . . . . . . . . . . . . . . . . . 14
3. Spray Card Cluster and Magnesium-Oxide Slides
. . . . . . 14
4.
EPA Meteorological Station. . . . . . . . . . . . . . . .15
5. Anemometer, Wind Vane, and 2-m Temperature Probe. . . . .15
6. Worker Mixing Desiccant. . . . . . . . ..
. . . . .
. . . 17
7.
Inspector Samples Desiccant Mixture. . .
. . . .
. . . . 19
. . 19
8. Tractor-Towed Trumbull Spray Rig. . . .
. . . . . . .
9. Disposal of Empty Desiccant Containers by Burning
. . . . 24
"10.
Empty Pesticide Containers at Northwest End of Airstrip. . 24
12. Same Location 3 Days After Application.
. . . . . . .
~ . . . 26
. . 26
11. Cotton Before Desiccant Application
. . . . . . .
13. Cotton Desiccated With Approximately 4.93% Arsenic Acid
(72 hours after application) ~ . . . . . . . . . . . . . .27

14. Cotton Desiccated with Approximately 9.74% Arsenic Acid
(72 hours after application) . . . . . . . . . . . . . . .27
i.v

-------
1.
INTRODUCTION
. Annually, millions of pounds of arsenic acid are used by farmers in
the United States to desiccate cotton prior to harvest. Arseni~ i~ an
environmentally persistent, toxic material, and is a suspected carcinogen. 1,2
The Environmental Protection Agency (EPA) is concerned about the continuirig .
and extensive use of arsenic acid as a preharvest cotton desiccant.
To monitor the effects of arsenic acid use, the National Enforcement
Investigations Center (NEIC) and EPA Region VI conducted a desiccant use
observation study. This study was accomplished at a cotton farm approximately.
30 miles south of Dallas, Texas, from September 22, through 28, 1976.
The objectives of the study were as follows: . .
1.
Ascertain through on-site observations whether the use of the
desiccant, arsenic acid, on cotton was in accordance with
label requirements and appropriate regulations.
2.
Document the environmental fate of an arsenical desiccant.
,3.
Identify use observation techniques of value to the EPA and to
other pesticide regulatory agencies.
The study determined residue levels of arsenic in ambient air,
vegetation, soil, water and fish samples collected before or after
desiccant application. Storage, handling, mixing and application of the
desiccant were observed. Weather conditions were monitored during and, '

-------
2
after the application. Various types of spray droplet cards and high- .
and low-volume. air sampling devices were used to characterize spray
drift.

-------
II.
SUMMARY AND CONCLUSIONS
In September 1976, the NEIC and Region VI conducted a 7-day study
in Ellis County, Texas to monitor the ground application of arsenic acid
used as a preharvest cotton desiccant. On-site observations, combined
with the use of various monitoring techniques, enabled the EPA to
determine user compliance with label instructions and to evaluate both
beneficial and harmful effects associated with the chemical desiccation
of cotton.
GENERAL CONCLUSIONS
On-site observations revealed that the applicator used a registered
chemical desiccant in a manner inconsistent with labeling and Federal
Regulations.
Environmental sampling showed that cropland, pasture and nearby
ponds were contaminated by the use of.arsenic acid.
The most valuable techniques used to document the practices and
(~nvironmental effects of the ground application were on-site evaluations
by trained observers and analyses of air, water and soil to determine
arsenic residues at the study site.
SPECIFIC CONCLUSIONS
1.
Pre-appl ication sampl ing revealed arsenic residues in soils
from the study site as high as 20 ~g/g. Additionally, five

-------
4
nearby ponds contained 9 to 870 ~g/l arsenic. Of
the two most contaminated ponds, one containing
about 50 ~g/l arsenic was used for livestock watering.
The water of neither of the ponds is safe for use as
a domestic or livestock supply.
2.
During mixing and loading operations, human health hazards
were observed. Workers handling the desiccant were not wearing
adequate safety apparel. Arsenic contamination in the mixing
area was of further concern because a residence and livestock
area were located within 100 m.
3.
Chemical analysis showed that one of five ba.tches of desiccant
was diluted improperly. This mixture contained half
the concentration of arsenic acid recommended on the desiccant
1 abe 1 .
4.
During the ground application, additional practices were
observed that were inconsistent with the desiccant label
instructions. Application was at a rate of 12 gal/acre rather
than the rate of 5 gal/acre; therefore, four of the five
batches introduced more than twice the arsenic onto the field
than was recommended. Application occurred when the wind
gusted to 15 mph, thus increasing off-field drift. The applicator
did not use an approved respirator as recommended on the desiccant
1 abe 1 .
5.
Observations by NEIC and Region VI personnel, after application
and prior to harvest, revealed that an application of 9.5%
arsenic acid, as recommended on the label, enhanced harvesting.
It minimized regrowth and produced a more brittle, non-pliable
foliage than did the lower concentration which was half the
recommended concentration.

-------
5
6.
During and following application, as much as 1,600 ~g arsenic
was found in the ambient air collected 90 m (300 ft) downwind
from the treated field.
-r
I.
Clean-up and disposal practices were inconsistent with the
label instructions and with Federal Insecticide, Fungicide and
Rodenticide Act (FIFRA), as amended, Regulations. Llquid
wastes were flushed from spray equipment o~~o the ground.
instead of being disposed of in a pit, as directed on the
label. Additionally, unrinsed plastic jugs previously containing
arsenic acid were burned on-site rather than being buried with

. . ,
the rinse water or encapsulated and buried in a,specially .
designated landfill, as designated in the regulations promulgated
pursuant to authority in FIFRA (40 CFR 165).
8.
The arsenic residues on treated foliage and cotton fiber were
as high as 46 ~g/g. These levels pose a health hazard to
wildlife, stock or humans entering the treated field: Further-
more, such levels could pose a health hazard to gin and mill
workers who handle the cotton on a day-to-day basis.
9.
Post-application sampling revealed an increase in arsenic
. .
levels in vegetation and water samples collected from fields.
. .

bordering the treated cotton. Increased arsenic levels in
these neighboring areas appeared to be the .result of spray
drift and contaminated runoff from the target field. Arsenic
contaminate~ bo~h'the foliage of nearby pastures and ponds
used for stock watering.
10.
Analysis of tadpoles obtained from a pond (Station 5) rev~aled

.. .
an arsenic level of 86 ~gjg. This indicated that atseniG

-------
6
accumulation in the aquatic food chain had occurred. Although
arsenic was not detected in fish, the potential for this
contamination exists.
11.
On-site observations and ,residue analysis for arsenic in air,
soil, water, and biota were the most valuable techniques for
evaluating the use and impact of arsenic acid. Spray-droplet
cards and slides were of limited value because of the sparsity
of droplet impressions collected. Tracer-dye studies were
unsuccessful because the fluorescence of the dye apparently
was quenched by the acidic properties of the desiccant or
because the dye concentration was too low.

-------
III.
DESCRIPTION OF STUDY AREA.
Ellis County is in northeast Texas approximately 30 miles south of
. .
Dallas and Fort Worth. Gently rolling grasslands with numerous small
ponds characterize the area. Although the land is used primarily for
pasture, thousands of acres are cultivated in cotton, sorghum, soybeans,
vegetables, and grains.
The site selected for the desiccant use observation study was a 79
ha (196 acre) cotton field in Ellis County, 4 miles west of Palmer,
Texas [Fig. 1J. The field wai sloped with the highest point toward its
nOl~thern edge. Surface runoff drained from the cotton field, across
adjoining pastures, and into nearby ponds or small creeks.
The ponds at Stations 1, 2, 3,4,7 and 8 are used primarily for
stock watering. The pond at Station 5 is used by the applicator as a
supply for diluting pesticides and desiccants. The pond at Station 6 is
fenced, preventing its use for livestock watering.
On the southwest side of the cotton field is a group of farm
buildings including a residence and a barn. The barn serves as a storage
shelter for small amounts of fertilizers, chemicals, and various farming.
implements. The area to the east and south of the barn is used for.
mixing and loading chemicals for ground application equipment. An
ab,:indoned airstrip bisects the field from near the residential area
toward the northwest. Mixing and loading of chemicals for aerial app1icatio~s
once occurred at the northwest end of the airstrip..
. Twenty-nine sampling stations were established in the study area.
These sampling sites were located in areas believed to have a high

-------
8
   ~ 
   -N- 
  ~ 1 
 ~  
  16  
 ~  Pasture 
Pasture 15  
23  12  
  14 10 ~
 Cotto n   
    9
29
Mix Area
Pasture
17
P astu re
~
20
LEGEND
POND
~
G)
Figure J. Study Area and Station Locations
Ellis County, Texas

-------
9
potential for receiving drift or surface water runoff from the treated
cotton field. Three stations were located in the cotton field; seventeen
in surrounding pasture land; eight in nearby ponds; and one near the
. .

farm buildings on the southwest side of the field [Fig. lJ. Air,
vegetation, soil, water and aquatic biota were collected at selected
times and locations in this sampling network [Table lJ. Samples were
analyzed to determine the environmental fate of arsenic sprayed on the
field to desiccate the cotton plants.

-------
10
Tab le. 1
SAMPLING STATIONS
ELLIS COUNTY, TEXAS
September 19'16
Station
Number
Water
.Soil
Vegetation
Drop 1 et 1 MgQ
Cards Slides
Greenburg-
Smith Impinger
High-Volume
Air Samplers
Aquatic
Biota
1 *     
2 *     
3 *     *
4 *     
5 *     *
6 *     
7 *     
8 *     
9   * *  
10  *   * 
11   *. *  
12   * *  
13  *    
14  *    
15   * *  
16   *   
17   * *  
18 *     
19   * *  
20  *   * 
21   * *  
22   * *  
23  *   * 
24   * *  
25  * * * * 
26   * *  
27  * * *  
28   * *  
29   *   
1 Mylar sheets, Thermofa:c~ Linagraph

-------
IV.
USE OBSERVATIONS
PRE-APPLICATION
~ampling
On September 24, environmental samples were collected to establish

background arsenic levels prior to ground app1icaion of arsenic acid

[Table 2J. Soil from the cotton field contained water-soluble arsenic

residues ranging from 0.75 to 0.95 ~g/g~ while soil from the mixing

and loading site (Station 25) contained 16 ~g/g.
Arsenic was not found in the cotton from the study field or in
grasses from nearby pastures. However, vegetation at the mixing and
loading site (Station 25) contained 20 ~g/g arsenic. The presence of
arsenic in the cotton field soils and in the mixing area indicated the
persistence of arsenical desiccants used by the farmer in previous
growing seasons.
Five of eight ponds near the cotton field also contained arsenic.
A pond north of.the study field contained 9 ~g/l arsenic, while the four
ponds to the south had from 9 to 870 ~g/l. The two ponds having the
highest amount of arsenic contamination (49 and 870 ~g/l) were either
adjacent to, or drained, the mixing and loading area. According to the
1976 QuaZity Criteria for water, published by the EPA, water is unsafe
as a domestic or livestock watering supply when arsenic concentrations
exceed 50 ~g/1.3 Therefore, the use of these two ponds for farming
activities is limited. Previous use of arsenic compounds on the farm
has contaminated cropland and nearby bodies of water.

-------
12
  Table 2     
  ARSENIC IN THE ENVIRONMENT     
  ELLIS COUNTY. TEXAS     
  September 19'16     
Station     Arsenic Residue
Number Type Sam pI e * Pre-Application Application Post-Application
1 Water   NO'"     NO
2 Water   NO     NO
3 Water   9     NO
4 Water   NO     , 9
5 Water   49     51
6 Water   870     890
7 Water   9     16
8 Wa ter   15     ' 20
18 Water        ,6l
10 Soil   NO     ND
13 Soil   0.95     106
14 5011   0.75. --   2.0
20 So 11    ND -...'   ND
23 Soil   ND     ND"
25 Soil   16     16
27 Soil   NO     NO
10 Vegetation ND     5.6
13 Vegetation ND     15
14 Vegetation ND     ' 46
20 Vegetation ND     ND
23 Vegetation ND --   8;5
25 Vegetation 20     5~'9
27 Vegetation ND     ND
3 Sunfish      ND
3 Bass        NO
3 Crappi e  ._- .   ND
5 Small fish  --, :  ND
5 Tadpoles      . 86
     ."",".':
10 High-Volume Air Sampler     ) 53
20 High-Volume Air Sampler     ) 1,300
23 High-Volume Air Sampler     >:1,600,
25 High-Volume Air Sampler      43
10 Greenburg-Smith Impi nger System ND 0.002', '  0.002'
20 Greenburg-Smith Impinger System NO 0.002   Np .
23 Greenburg-Smith Impinger System NO NO . .   ND
25 Greenburg-Smith Impinger System NO 0.002 '   0.001
II Val!lss expresosd:
.. ND = not detected.
water - ug/l; air - ug; all other samples ug(g.

-------
13
Prior to observing mixing and loading operations~ sampling devices

. .
wer~ placed around the cotton field to evaluate drift of the desiccant
during ground application. These devices included: Greenburg-Smith
impinger units; high-volume air samplers~ spray-droplet cards (Linagraph
and Thermofax paper); magnesium-oxide-coated glass slides, and mylar
sheets [Figs. 2, 3]. Each devic~ is described in the Appendix of this
report.
The Greenburg-Smith impinger units were operated at four locations
(Stations 10~ 20, 23 and 25) several hours before mixing, loading
and applying the arsenical desiccant. The reference air sample collected
by the impingers revealed no detectable arsenic «0.001 ~g of arsenic)
in the air during this pre-application period [Table 2].
A meteorological station was assembled on the property near the
mixing site [Figs. 4, 5], to measure and record wind direction and
velCicity~ air temperature, lapse rates and precipitation. These factors
were recorded during and following the chemical treatment of the cotton
field; results were used to evaluate the influence of ~tmospheric conditions
on the translocation of arsenic beyond the target field.
Mixing and Loading Activities
On the morning of September 25~ farm workers began the mixing and
loading of the chemical. Three lots~ containing 175 one-gallon plastic
jugs of Hi-Yield .R H-10,* were stack~d neatly on the ground beside a
tractor and spraying rig. An acid-resistant 55 gal (208 liter) stainless
steel drum was used to premix the desiccant. The applicator hoped to
demonstrate to the EPA that the concentration specified on the label was'
ineffective for preharvest cotton desiccation. The applicator planned'
* lli-Yield R H-10 EPA Reg. No. 7401-195.

-------
14
::!,---
_.
. ,
. ,
Figure 2.
card
~;
s~ V

Greenburg-Smith impinger
platform, High Volume Air
unit, dropl et
Sampler.
Figure 3.
Spray droplet card cluster and
magnesium oxide slides.

-------
15
.1
~1


~ ..
I i
Figure 4.
EPA Meteorological Station.
p:. -;;- '.- '-II.... ?,'
"~~ ','
n " ~o 1t~..
r
,.,- ~ ..
(
~
-- ~Irl--
~
I
L--. - --
-
Figure 5.
Anemometer, wind vane, and 2-meter
temperature probe.

-------
16
to mix one batch at recommended label 'concentration, and the four
remaining batches at a higher concentration.
The low concentration batch was mixed by pouring 15 gal of Hi-Yield
H-10 and 1 pt wetting agent into a stainless steel.drum.. Two jugs of..
the concentrated desiccant were poured simultaneously so. their flow
streams intermingled and splashing was minimizedIFig. 6]. The chemical
mixture was pumped into the spray tank and water was added to increase
the volume to approximately 500 gal. The high c~ncentration batch.was
mixed by increasing the volume of 40 gal Hi-Yield'H-1Q arid 1 qt wetting
agent to approximately 500 gal.
Chemical analysis confirmed that the low concentration batch was
diluted to half the label recommendation, and the other four
batches had arsenic concentrations within the label recommendation "
[Table 3].
At the request of the EPA investigators a fluorescent tracer dye
(Rhodamine WT) was added into each batch in the tractor-'moUnted spray
tank. The use of this dye will be discussed later. During the mixing
and loading operati'on, farm work~rs were careful not to sp1.ll the chemical,:
mixt~re on their clothing, the tractor rig or the ~~o~nd~, In;addition
to regular clothing, workers wore rubber gloves~ Work crew~ did not
wE!ar.respirators, goggles or coveralls. Protective'clQthing worn by EPA
observers consisted of long-sleeved coveralls, gloves, bO,ots, face.
shields, hats, rubber apron, and approved canister-type respiratqrs
[Fig. 7]. Protective clothing was worn by the EPAteam both 'during an~:
after application. -' '. .

-------
17
, .,. "
,~- ~'-
",.~'.... ".'1...:: ..,,,'.
~.=-j,£... ,-
- / r...
, ~
Figure 6.
Worker mixing desiccant.

-------
Brand
Name
Hi-Yield
H-10
Tab le :5
Formulation and Diluted MateriaZ Analysis
(% by Weight)
Desiccant Concentrate
Arsenic Content
Indicated on Label
Actual Arsenic
Content
Arsenic Acid Content
Indicated on Label
Batch
Number
1
2
3
4
5
39.61%
41.61%
75.00%
Use Dilution Preparation
% Arsenic
% Arsenic Acid
4.89
5.76
2.60
5.14
5.09
9.26
9.02
4.93
9.74
9.64
18
Actual Arsenic
Acid Content
78.84%

-------
1-
19
Figure 7.
Inspector samples desiccant mixture.
Figure 8.
Tractor-towed Trumbull Spray Rig.

-------
20
APPLICATION
On-Site Observation
On September 25, between, 10:30 a.m. and 7:30 p.m., arsenic acid
dE!siccant was applied to the study field. Application was delayed
because heavy dew, which would reduce the effectiveness of arsenic acid,
covered the field. Five loads (approximately 2,500 gal) of diluted Hi-
Yield H-10 and wetting agent were sprayed on the cotton with a tractor-
drawn rig.
Warm and sunny weather prevailed during application [Table 4J. Air
temperature ranged from 77 to gO°F, and no inversions were detected.
Winds were variable, predominantly from the south at 3 to 8 kmh (2 to 5
mph) with occasional gusts to 24 kmh (15 mph).
The Trumbull Spray Rig [Fig. 8J was used to apply the desiccant.
The tractor-towed rig consisted of three spray booms; the center boom
was mounted permanently to the frame while the two lateral booms were
movable. The booms were equipped with a total of 25 spray nozzles (03-
45), evenly spaced, 51 cm (20 in) apart.
The three booms were preset to a height of 1.2 m (4 ft) above the
qround. This position allowed a few centimeters' clearance between the
boom and the top of the cotton plants. The desiccant was sprayed at
.L 23 kgjcm2 (6.0 ps i) uniformly over the pl ants at a r'ate of 12 gal jacre
(112.1iterjha). A twelve-row swath of cotton was treated in a single
pass.
As discussed previously, two concentrations of desiccant were
applied to demonstrate that the recommended mixture was ineffective.

-------
      21'
   Table 4    
 Weather Conditions During Application  
   WIND  AIR TEMPERATURE
Batch Time of Speed Di rect i on IAJration 10 m 2.m Lapse *
Number Application mph  minutes of . uF Rate
  <1 to 8 SW 8.0 77 to 81 77 to 82 0.5
 1030-1200  S 74.0   (lapse)
   SE 8.0   
  5 to 11 SW 7.8 83 to 86 83 to 87 0.5
2 1230-1400  S 71.0   (1 a ps'e )
   SE 11 .2   
  2 to 1 5 SW 13.0 85 to 86 86 to 88 1.5
3 1410-1530  S 62.2   (1 apse)
   SE 4.8   
  4 to 1 3. SW 5.7 86 to 87 87 to 90 2.0
4 1545-1750  S 99.7   (1 apse)
   SE 19.6   
5 1810-1930 1 to 3 S 55.4 84 to 86 . 83 to 87 0
   SE 24.6  (normal)
* A numerical classification relating to air stability.

-------
22
Four batches were prepared in accordance with the label which .states
"Mixl to 1 1/2 qt with sufficient water to make 5 gal of spray and
apply to each acre of cotton. II
The high concentration of desiccant was applied to approximately 64
ha (157 acres) of cotton, while the lower concentration was sprayed on
the l"emaining 16 ha (39 acres). The low concentration batch was prepared
at half the recommended concentration (1/2 qt/5 gal).
Since these mixtures were applied to the cotton at the rate of 12
gal/acre rather than the recommended rate of 5 gal/acre, application of
the four batches mixed at the proper concentration resulted in more than
twice the recommended amount of arsenic acid being applied to each acre
of cotton.
Although the improperly diluted low concentration batch resulted
in applying the recommended amount of arsenic acid to each acre of
cotton, the mixture was too dilute to desiccate the cotton effectively.
The EPA observation team reported no visible spray mist beyond the
tal~get field. The tractor operator turned off the spray pump when he
left the field at the end of each treated row. This technique reduced
off-field contamination 'of grass lands (pasture) and soils. The applicator
was not observed to be wearing a respirator to protect him from spray
drift, as directed on the label. However, he was enclosed in a tractor cab;
the degree of protection afforded was not evaluated but should be in future
use studies.
P'OST -APPLICATION
~ixing Area
Upon completing the desiccant treatment, the applicator drove
tractor and spray-rig to the mixing area to be cleaned. Clean-up

-------
23
proc1:!dures consisted of mixing approximately 2.7 kg (6 lb) of sodium
bicarbonate with about 250 gal of water in the spray tank. Sodium
bicarbonate neutralizes acidity but may not degrade arsenic. The
mixture was discharged through the nozzles at the edge oJ the treated
cotton field (about 35 yd from the mixing area). A dike around a small
pond (Station 5) adjacent to the rinse-water dump site prevented 'runoff
- from the cleaning area into the pond. However~ surface water drainage
from this area was to the southwest toward another nearby pond (Station
6). This clean-up procedure was not in compliance with instructions on
the Hi-Yield H-10 label which states "Rinse spray equipment and containers
and dispose of liquid wastes in a pit in non-cropland located away from
water supplies."
The following day~ the empty plastic jugs and carboard boxes that
previously contained the desiccant were burned [Fig. g). The EPA specifically
recommends against the open burning of containers formerly used for
arsenic compounds. The regulation (40 eFR 165.9) recommends that arsenic
containers be triple-rinsed~ punctured and buried with the rinsings.
Otherwise~ the unrinsed container should be encapsulated and buried in a
specially designated landfill.
'In addition to this improper procedure for disposing of empty
desiccant containers~ the applicator has the further problem of disposing
of empty pesticide containers. Currently~ no landfill is available to
local farmers and applicators for the disposal of used chemical containers.
Figure 10 is a photograph of the applicator's aircraft loading zone a~
the northwest end of his airstrip. The applicator stated that he hasn't
used this airstrip for several years. Empty pesticide containers and
general litter at the loading zone illustrates the ,magnitude of this'
applicator's disposal problem.

-------
24
{..~;.
~
1::1 Q
O...~b
Figure 9.
Disposal of empty desiccant containers
by burning.
Figure

-------
i -~-
I
~~-- -.--..--- --------~ - -. -.
25
I

,

I
Soil and vegetation samples from the mixing site (Station 25) were
analzyed for arsenic residues. The soil contained an arsenic concentration
of 16 ~g/g, while vegetation contained ,as m~ch as 5.9 ~g/g [Table 2]. .
This environmental contamination was of concern because a residence and
a livestock area were within 100 meters of the arsenic mixing site.
Although adverse health effects were not observed, the potential was
high for arsenic contamination in food or water consumed by humans and
farm animals that occupied the neighboring area.
Cotton Field
Within hours of being sprayed, the cotton leaves were wilting and
turning brown. Destruction of the foliage appeared to be most rapid in
'. .

portions of the field treated with the higher concentration of arsenic
acid. Figures 11 and 12, photographed at the same location before and
. .
after application, illustrate the efficacy of desiccation with arsenic
acid. .
Three days after the desiccant application,' the Region VI Consumer
Safety Officer photographed areas of the cotton field which had been
treated with either high or low concentrations of arsenic acid. The
mixture containing approximately 5% arsenic acid allowed some leaves to
I"emain healthy (green and sappy) [Fig. 13]. Conversely, the mixture
containing about 9.5% arsenic acid appeared to desiccate the cotton
plants complete11. The higher concentration produced brittle, non-
pliable foliage [Fig. 14].
Local cotton growers strive for ~ degree of desiccation that
produces brittle plants and minimizes leaf regrowth. These conditions
enhance harvesting efficiency and permit earlier harvesting. Consequently,

-------
26
Figure 11.
Cotton before desiccant application.
Figure 12.
Same location 3 days after application.

-------
27
Figure
Figure 14. Cotton desiccated with 9.74% Arsenic
Acid (72 hours after application).

-------
28.
local growers use 9.5% arsenic acid as a preharvest cotton desiccant.
This is the recommended concentration. However, the applicator observed
. .

in this study wasted arsenic acid by applying at more than double the
recommended rate.
As discussed earlier, cotton plants and soil from the study field
\~ere analyzed for arsenic residue prior to the desiccant treatment. The
plants contained no measurable amount of arsenic. However, the soil in
which the cotton was grown contained as much as 0.95 ~g/g arsenic [Table
2J. Apparently, uptake of measurable quantities of arsenic from the
soil and storage in the cotton plants did not occur.
After the cotton crop was chemically desiccated, soil and cotton
were again collected from the treated field for arsenic analysis. The
samples were collected from two areas: one treated with 5% arsenic acid
(Station 13); and the other treated with 9.5% (Station 14). Arsenic
concentrations in soils from these areas were 1.6 and 2.0 ~g/g, respectively,
.about 70 to 170% increase. Arsenic concentration in the cotton plants
contained 15 to 46 ~g/g arsenic, respectively, in the 5 and 9.5% treated
areas [Table 2J.
These arsenic residues in soils and cotton are significant because
of their relationship to environmental and human health hazards. Some
of the important issues are illustrated by precautions on the desiccant
label stating:
IIHi-Yield H-10 containing T-49.is a very poisonous material and
should not be applied under any circumstances to cotton that will
be harvested by humans.

-------
29
"Avoid contamination of food and water to be consumed by humans or
animals.
"00 not graze or use treated plants or gin trash for feed or
forage. Keep livestock, poultry and pets off of treated areas.
"This product will kill fish and wildlife. Do not contaminate
streams, lakes, ponds, woodlands or other non-crop areas. Birds
and other wildlife in treated areas may be killed. Apply this
product only as specified on this label. II
Another significant issue related to arsenic residue on cotton is
that it could pose a health hazard to: (1) gin and mill workers who handle
th€~ raw cotton on a day-to-day basis; (2) persons using cotton seed oil for
food; and (3) livestock consuming cotton seed meal.
Surrounding Areas
Dr~: ft
Following the completion of the desiccant application, EPA teams
entered the study site (target field and surrounding area) to retrieve
spray-droplet cards and slides. Spray-droplet images were readily
visible on Thermofax cards and glass slides at two locations (Stations
. .

16 and 29) each 90 m (300 ft) north and west of the field, and at
numerous other locations surrounding the field [Table 5]. Droplets were
not det~cted on the Linagraph cards. To obtain a Volume Me~n Diameter
. .

(VMD)* of spray droplet size, a minimum of 200 droplet impinge~ent
craters. must be measured on magnesium-oxide-coated slides. None of the
slides exposed during the desiccant application showed a sufficient
* VMD is that volume which divides the droplet diameter into two equal
parts, one-half above and one-half below the median or.50% cumulative
p01:.J~.t -

-------
     30
  TabZe 5  
 DropZet Impingements on ThePimofax Paper 
 Di stance - Direction  
Station from from Wind Direction  
Number Cotton Field ai1Q Speed  Drop1ets/cm2
9 30M (100ft) E from SE to SW, 4-9mph 22
11 30M (100ft) N from SE to SW, 4-9mph 16
'12 30M- (100ft) N from SE to SW, 4-9mph 11
"15 30M (100ft) N from SE to SW, 3-7mph 37
'16 90M (300ft) N from SE to SW, 3-7mph 5
- 17 30M (1 OOft) S from SE to SW, 1-3mph 10
19 30M (100ft) N from SE toSW, 3-7mph 12
21 30M (100ft) S from SE to SW, 1-3mph 17
22 30M (100ft) N from SE to SW, 3-1mph 16
24 30M (100ft) S from SE to SW, 1-3mph. 32
25 30M (1 OOft) S from SE to SW, 8-15mph. 24 -
26 30M (100ft) N from SE to SW, 3-7mpb - 10
27 30M (100ft) S from SE to SW, 8-15mph 22
28 30M (100ft) W from SE to SW, '8-15mpb 25
29 90M (300ft) W from SE to SW, 8-15mph 22

-------
I
31
number of craters to yield a VMD. Therefor~, valid estimates of spray
drift potential, based on droplet size, could not be made.
.
During application, air samples for arsenic analysis were collected
by means of Greenburg-Smith impinger units in pastures bordering the
treated cotton (Stations 10, 20, 23, and 25); each location was approximately
'90 m (300 ft) from the target field. Air collected to the east and
sDuih (Stations 10 and 20), and at the mixing site (Station 25), contained
0.002 ~g arsenic [Table 2].
The Greenburg-Smith impingers were set to collect a continuous,
two-hour, post-app1icati~n air sample. As these units were manually
. ,
s\~itched from one impinger to another during pre-application, applicaton
and post-application, the air inlet tubes of the impingers, not in use,
ware capped with desiccant-free aluminum foil to prevent contamination
. .

of the impingers.
Analysis of Greenburg-Smith impinger samples confirmed that arsenic
remained in the ambient air after treatment of the cotton field [Table
2]. The impinger unit at Station 10, approximately 90 m (300 ft) northeast
from the treated field, collected 0.002 ~g arsenic, and 0.001 ~g arsenic
, '
was collected at t~e mixing and loading site, Station 25.
High-volume air samplers, operated at four locations around the
target field (Stations 10, 20, 23, and 25), collected a compositR air
sample ~uring and following the desiccant application. These sampling
devices collected arsenic at all four air sampling stations. On-site
meteorological information showed that during the desiccant application
the wind gusted to 15 mph from the south across the target field toward
Station 23. The high-volume filtering unit operated in the northern
. pasture at Station 23 collected the largest amo~nt of arsenic drift,
1 ,600 ~g. The evening of September 25, after the desiccant application,

-------
32
the wind direction changed. The wind blew from the north across the
treated field toward a pasture located southeast. The filtering unit
operated in this pasture, approximately 90 m (300 ft) downwind from the
treated field, collected 1,300 ~g of arsenic from the ambient air.
Other air filtering units operated in the pasture northeast of the
treated field (Station 10) and at the mixing area (Station 25) collected
53 and 43 ~g of arsenic, respectively. These results showed that arsenic
drifted into nearby fields a distance of 90 m (300 ft) or more.
The fluorescent tracer dye (Rhodamine WT) added to the desiccant
and sprayed on the cotton was not detected in air samples collected by
the Greenburg-Smith impinger units or by the high-volume samplers.
Furthermore, spray droplets rinsed from mylar sheets, placed near the
treated field, produced no measurable fluorescence. Because droplet
cards and slides indicated drift, it was expected -that the dye study
should have also; however, no drift was detected by this means. Lack of
fluorescent response may be attributable to the acidic properties of Hi-
Yield H-10 destroying the fluorescent quality of the tracer dye. It is
also highly probable that the dye concentration was undetectably low.
Prior to the application, arsenic was not detected on vegetation
from pastures surrounding the cotton field [Table 2J; arsenic was present
on vegetation at the mixing site (Station 25) both before and after the
application. Vegetation collected after application from pastures to
the northeast and north (Stations 10 and 23) had 5.6 and 8.5 ~g/g arsenic,
respectively, while arsenic was not detected on the vegetation of pastures
to the south or west. The contaminated vegetation was from pastures
shown (by analysis of spray cards, Greenburg-Smith impingers, and high- .
volume air samplers) to have received dri~t from the arsenic acid application.
Additionally, analyses revealed that the soils of pastures surrounding
the cotton field did not contain water-soluble arsenic, either before or
after the application [Table 2J. Therefore, it is concluded that the

-------
~
33
arsenic contamination of pasture vegetation to the northeast and north
of the cotton field resulted from drift, rather than uptake from the
soil. Cattle observed grazing this vegetation would accumulate arsenic
. .
and could become sensitized to it, becoming more vulnerable to poisoning
from future arsenic exposures.
Surface drainage collected after application from Station 18 on the
southeast side of the field contained 61 ~g/l arsenic. Analysis of
water from nearby ponds showed a pattern of increased arsenic contamination.
Py'ior to treating the cotton field with arsenic acid, the three ponds
south of the target field and the pond at the mixing area contained 9 to
870 ~g/l arsenic. Following the cotton desiccant treatment, arsenic
concentrations in the four ponds generally increased to a range of 16 to
890 ~g/l [Table 2]. The most contaminated pond, Station 6, was located
in the area draining the chemical-waste dump site. No control measures,
such as diking, have been taken to prevent contamination of the pond
(Station 6) by surface runoff from the dump site.
The maximum concentration of arsenic permissible in water used for
liv~stock is 50 ~g/1.3 The pond in the mixing area contained about 50 ~g/l,
and was used for livestock watering. The most contaminated pond (Station
6) was located in a wooded area which was fenced. The pond and adjacent
woodland appeared to be a wildlife habitat but was not used for livestock
grazing and watering.
On Septembe~ 27, fish and tadpoles were collected from the ponds at
Stations 3 and 5. Although none of the fish samples contained arsenic,
tadpoles from the pond in the mixing area contained 86 ~g/g [Table 2J.
Thus, arsenic accumulated in the aquatic food chain, and arsenic contamination
in fish could occur in the future.
To summarize, arsenic from the desiccant application left the
cotton field by drift and surface runoff. This arsenic contaminated
both the foliage of nearby pasture lands and ponds used for livestock
watering.

-------
V.
EVALUATION OF METHODS
Many of the techniques used by the EPA to monitor aerial applications
of pesticides were less effective in monitoring ground applications.
Nevertheless, two techniques proved useful: 1) on-site observations,
and 2) arsenic-residue analysis of environmental samples collected from
the study site. An evaluation of the techniques used in this study is
presented below.
ON-SITE OBSERVATIONS
Visual observations provided first-hand information about the use
practices of a preharvest cotton desiccant, arsenic acid. On-site
visits also afforded the applicator an opportunity to discuss with EPA
officials some of the agricultural problems caused by current pesticide
and chemical-use regulations.
AIR SAMPLING DEVICES
High-volume air samplers and Greenburg-Smith impinger units captured'
measurable amounts of arsenic carried by wind currents off the target
field. The high-volume air sampler appeared to be the more efficient'
device possibly because it sampled a larger volume of air or because'the
dry fiberglass filter used in it may have a better trapping efficiency
than the 1% sodium hydroxide solution used in the Greenburg-Smith impinger
device.

-------
35
TRACER-DYE STUDIES
The dye studies performed to evaluate drift characteristics were
unsuccessful. Fluorescent dye (Rhodamine WT) was not detectable after
being added to the cotton desiccant mixture and sprayed on the field.
It is believed that the fluorescence of the dye was quenched by the
acidic properties of the desiccant (arsenic acid), or that the dye
concentrations were undetectably low.
SPRAY-DROPLET CARDS AND SLIDES
Spray drift was detected successfully by droplet impressions on
Thermofax cards and on glass slides coated with magnesium oxide.
Droplets were not detected on Linagraph cards.
Ground application apparently minimized
small quantities of droplets impacted on the
droplet impressions precluded size analysis.
spray drift resulting in
slides. This sparsity of
ENVIRONMENTAL SAMPLING
One of the most useful monitoring techniques in this study involved
pre- and post-application environmental sampling. Residue analysis for
arsenic showed that application of the preharvest cotton desiccant
contaminated adjacent fields and nearby surface waters.

-------
36
REFERENCES
1.
Buchanan, W. D., 1962. Toxicity of Arsenic Compounds.
Elsevier Publishing Co., New York, 155 pp.

Sommers, S. C. and R. A. McManus, 1953. "Multiple Arsenical
Cancers of the Skin and Internal Organs." Cancer, 6:347.
2.
3.
Quality Criteria for Water. U. S. Environmental Protection
Agency, Washington, D. C., 1976. p. 25.
4.
Standard Methods for the Examination of Water and Wastewater.
American Public Health Association, Washington, D. C., 14th Ed.,
1975. p 283 ff.
5.
AOAC Methods of Analysis. William Horwitz, Ed. Association of
Official Analytical Chemists, Washington, D. C., Eleventh Ed.,
1 970. P 399 ff.

-------
Appendix
Sampling Devices and Methods

-------
SAMPLING DEVICES AND METHODS
Greenburg-Smith Impinger
This system is a semi-quantitative method for detecting airborne
desiccant levels. The system consists of a pump which draws ai,r'throug'h
an impinger and auxiliary equipment to control and measure air flow.
Each unit was numbered, assigned to a statiori and used only at that
station. Air-flow rates were set at 0.5 ft3/min. Switching the flow,
from one impinger to another was done manually to insure the proper
impinger was in operation during the pre-application, application and
post-application sampling periods. Desiccant-free aluminum foil was
. ,

used to cover the impinger air intake tubes before and after sample
collection. The impingers were filled with approximately 150 ml of 1%
sodium hydroxide solution.
To establish airborne arsenic levels around the field, the first
, ,

impinger was operated for two hours before desiccant application.
, ,

During application, a time span of approximately nine ho~rs, ,the secorid
impinger was in operation to capture desiccant drift. At the conclusion
of application, the third impinger was in operation for approximately
two hours to entrap residual desiccant that remained ai~b6~ne after
application. '
High-Volume Air Samplers
This system uses a fiberglass filter 20 x 25.4 cm, on which airborne
particles and spray are trapped. In use,appro~imat~ly1 m3/min

-------
".

r
39
(35 ft3/min) of air is drawn through the filter. The system was operated
for thirteen hours. A total of 772 m3 (27,300 ft3) of air passed through
each filter. The filters were removed, stored in plastic bags and
returned to the laboratory. The filters were prepared for analysis by
. eluting the entrapped material with 200 m1 distilled water. The e1utrate
was then analyzed for arsenic.
Magnesium-Oxide-Coated Slides
These slides were prepared by burning thin strips of magnesium
beneath the slide to form a white, powdery coating. On impact with the
stationary slide, the airborne droplet formed a crater which provided
visible evidence of droplet impingement. Craters thus formed can be
measured to within 5 microns at 100X .magnification. Provided a minimum
of two hundred craters can be measured on each slide, the Volume Mean
Diameter of spray droplet size can be computed and an estimation of
drift potential can be made.
Spray Droplet Cards
Spray card ~lusterswere constructed by stapling Linagraph 480
paper, Thermofax 209 copy, type 640 paper and mylar sheeting onto 30 cm
(12 inch) square poster-board base. The clusters were sited around the
field atop 1 meter upright wooden platforms and attached by elastic
bands.
Linagraph and Thermofax paper were extremely sensitive to water or
petroleum-based droplets. Visual images were formed on the paper upon
droplet contact.. Spray droplet cards of 10 x 14 cm mylar sheets were
used only to collect spray droplets for fluorescence analysis of tracer
dye material (Rhodamine WT). -

-------
40
The droplet card impingements were counted at NEIC using a 30X
binocular microscope. Cards with low numbers of impingements were
counted in entirety. On cards having high density impingements, five
random 2.5 cm (1 inch) square areas were counted and the average obtained.
Droplet impingement counts were expressed as the number of droplets per
2 .. .
cm. .
Fluorescence Analysis
Laboratory analyses for fluorescent response of tracer dye were
done using a Turner Model 111 fluorometer with a high sensitivity door.

. .
Sample material from the Greenburg-Smith impingers was analyzed and
compared with a 1% sodium hydroxide blank. Mylar sheets were washed
with 100 ml of 95% ethyl alcohol and fluorescent response compared with;
an ethyl alcohol blank.
Analytical Methods
Water samples were analzyed by the graphite furnace flameless
atomic absorption method. All other arsenic analyses were conducted
using the silver diethyldithiocarbamate method as described in Standard
Methods. 4
Organic samples (fish, vegetation) were dried at 105°C and either
leached or digested. For digestion, samples were ground in a blender
and an aliquot was placed in a micro-Kjeldahl flask and digested with
, . '. .

nitric and sulfuric acids according to AOAC preparation method number'
. . . .

25.008.5 Cotton boll samples were leached for two days with deionized
water; the water was then squeezed out and analyzed.
Soil samples were dried at 105°C and ground to less than 300 mesh,
in a rotary mill. Ten gram aliquots were then leached with water. The'

-------
41
supernatant liquid was centrifuged twenty minutes at 2,000 rpm to remove
solids and analyzed.
Minimum detection limits for the water sample analysis was <9 ~g/l.
Thf~ lower limit on the Greenburg-Smith impinger unit samples was <0.001 ~g.
The lower limit for organic samples was <4 ~g/g.

-------