II.
E.
2
­
Page
1
of
46
II.
E.
2
Residential
Exposure
Scenarios
Appendix
The
Preliminary
NMC
CRA
considered
a
variety
of
exposure
scenarios
for
consumer
applicator
and
post­
application
residential
exposures.
The
data
from
multiple
studies
that
measured
various
exposure
values
have
been
used
in
the
residential
portion
of
the
residential
risk
assessment.
In
some
cases,
statistical
distributions
have
been
fitted
to
the
datasets.
For
such
datasets,
exposure
estimates
were
based
on
the
fitted
distributions.
Brief
descriptions
of
the
studies
and
statistical
details
of
the
datasets
used
in
the
residential
portion
of
the
risk
assessment
are
provided
below.

1.
Lawn
Care
Exposure
Scenarios
Unit
Exposure
Data
Study
Summaries
MRID
449722­
01
(
ORETF
Turf
Handler
Studies):
A
report
was
submitted
by
the
ORETF
(
Outdoor
Residential
Exposure
Task
Force)
that
presented
data
in
which
the
application
of
various
products
used
on
turf
by
homeowners
and
lawncare
operators
(
LCOs)
was
monitored.
All
of
the
data
submitted
in
this
report
were
completed
in
a
series
of
studies.
The
two
studies
that
monitored
homeowner
exposure
using
a
granular
spreader
(
ORETF
Study
OMA003)
and
a
hose­
end
sprayer
(
ORETF
Study
OMA004)
are
summarized
below.

OMA003:
A
total
of
30
volunteer
test
subjects
were
monitored
using
passive
dosimetry
(
inner
and
outer
whole
body
dosimeters,
hand
washes,
face/
neck
wipes,
and
personal
inhalation
monitors).
Each
test
subject
carried,
loaded,
and
applied
two
25­
lb
bags
of
fertilizer
(
0.89%
active
ingredient)
with
a
rotary
type
spreader
to
a
lawn
(
a
turf
farm
in
North
Carolina)
covering
10,000
ft2
(
one
bag
to
each
of
the
two
5000
ft2
test
plots).
Application
to
each
subplot
continued
until
the
hopper
was
empty.
Each
participant
also
disposed
of
the
empty
bags
at
the
end
of
the
replicate.
The
target
application
rate
was
2
lb
ai/
acre
(
actual
rate
achieved
was
about
1.9
lb
ai/
acre).
The
average
application
time
was
22
minutes,
including
loading
the
rotary
push
spreader
and
disposing
of
the
empty
bags.
Approximately
0.45
lb
ai
was
handled
in
each
replicate.
Dermal
exposure
was
measured
using
inner
and
outer
whole
body
dosimeters,
hand
washes,
face/
neck
washes,
and
personal
air
monitoring
devices
with
OVS
tubes.
Overall,
residues
were
highest
on
the
upper
and
lower
leg
portions
of
the
dosimeters.

OMA004:
Dermal
and
inhalation
exposures
were
estimated
using
passive
dosimetry
techniques
(
biological
monitoring
data
were
not
collected).
A
total
of
60
replicates
were
monitored
using
30
test
subjects
(
two
replicates
each)
during
applications
to
residential
lawns
in
Frederick,
Maryland.
Thirty
applicator
replicates
were
monitored
using
a
ready­
to­
use
(
RTU)
product
(
Bug­
B­
Gon)
packaged
in
a
32
fl.
oz.
screw­
on
container.
These
containers
were
attached
to
II.
E.
2
­
Page
2
of
46
garden
hose­
ends.
An
additional
30
mixer/
loader/
applicator
replicates
were
monitored
using
Diazinon
Plus
also
packaged
in
32
fl.
oz.
plastic
bottles.
This
product
required
the
test
subjects
to
pour
the
product
into
dial­
type
sprayers
(
DTS)
that
were
attached
to
garden
hose­
ends.

A
nominal
application
rate
of
4
lb
ai/
acre
was
used
for
all
replicates.
Each
replicate
monitored
the
test
subject
treating
5,000
ft
2
of
turf
and
handling
a
total
of
0.5
lb
ai/
replicate.
The
average
time
per
replicate
was
75
minutes.
Dermal
and
inhalation
exposure
were
measured
using
inner
and
outer
whole
body
dosimeters
(
long
pants
and
long
sleeved
shirt
over
long
underwear),
hand
washes,
face/
neck
washes,
and
personal
air
monitoring
devices.

MRID
44459801
(
Applications
Of
Carbaryl
To
Vegetable
Gardens):
The
data
collected
reflect
the
dermal
and
respiratory
exposure
of
homeowners
mixing,
loading
and
applying
RP­
2
Liquid
(
21%),
a
carbaryl
end­
use
product.
Applications
were
made
by
volunteers
to
two
18
foot
rows
of
tomatoes
and
one
18
foot
row
of
cucumber.
The
only
test
field
was
located
in
Florida.
For
this
study,
RP­
2
Liquid
(
21%)
exposures
were
monitored
using
hose­
end
sprayers
and
low­
pressure
hand
wand
sprayers.
Exposures
to
Sevin
®
10
Dust,
using
a
separate
duster
device
that
required
transfer
from
the
package
and
Sevin
®
Ready
To
Use
Insect
Spray
(
RTU)
in
a
trigger
sprayer
package
were
also
monitored.
Exposure
for
each
spray
method/
product
combination
was
monitored
using
40
handlers
(
replicates).
Of
the
40
replicates
per
spray
method/
product
combination,
20
wore
household
latex
gloves
and
20
performed
tasks
without
gloves.
The
20
dust
product
replicates
loaded
the
dusters
and
applied
without
gloves
only.

Each
replicate
opened
the
end­
use
product,
added
it
to
the
application
implement
(
except
the
RTU
product),
adjusted
the
setting
and
applied
it
to
the
vegetable
rows.
After
application
to
the
vegetable
rows,
dosimeters
were
collected.
Inhalation
exposure
was
monitored
with
personal
air
sampling
pumps
with
OVS
tubes
attached
to
the
shirt
collar
in
the
breathing
zone.
Dermal
exposure
was
assessed
by
extraction
of
carbaryl
from
inner
and
outer
100
percent
cotton
dosimeters,
face/
neck
wipes,
and
glove
and
hand
washes.
The
inner
and
outer
dosimeters
were
segmented
into:
lower
and
upper
arms,
lower
and
upper
legs,
front
and
back
torso.

Dermal
exposure
was
determined
by
adding
the
values
from
the
bare
hand
rinses,
face/
neck
wipes
to
the
outer
dosimeter
lower
legs
and
lower
arms
plus
the
inner
dosimeter
front
and
rear
torso,
upper
legs,
lower
legs,
lower
arms,
and
upper
arms.
This
accounts
for
the
residential
handlers
with
barehands
wearing
short­
sleeved
shirt
and
short
pants.
II.
E.
2
­
Page
3
of
46
Statistical
Details
Distributional
parameters
were
estimated
for
the
dermal
and
inhalation
unit
exposure
(
UE)
values
for
the
granular
(
Table
II.
E.
2­
1),
dust
(
Table
II.
E.
2­
2),
and
liquid
sprayable
(
Table
II.
E.
2­
3)
formulations
of
Carbaryl.
All
dermal
and
inhalation
UE
values
represent
milligrams
exposure
per
pound
of
active
ingredient
of
a
pesticide
handled.
All
UEs
were
assumed
to
be
lognormally
distributed
(
i.
e.
fitted
with
a
lognormal
distribution).
For
each
dataset,
the
shape
(
 )
and
scale
(
 )
lognormal
parameters
were
estimated
by
calculating
the
mean
and
standard
deviation
of
the
natural
logarithms
(
base
e)
of
the
UEs.
Parametric
estimates
of
the
arithmetic
mean
(
µ
)
and
standard
deviation
(
 )
of
the
lognormal
distribution
were
then
calculated
based
on
the
shape
and
scale
parameter
estimates.
The
formulae
used
to
calculate
the
mean
and
standard
deviation
are
given
below.

)
 
exp(
 
µ
2
2
1
+
=

1
)
exp(
 
µ
 
2
 
=

Shapiro­
Wilk
(
S­
W)
normality
test
statistics
were
used
to
assess
the
lognormal
assumption
implicit
in
the
parametric
calculations
of
the
mean
and
standard
deviation.
The
means,
standard
deviations,
and
p­
values
of
the
S­
W
statistics
are
provided
in
Table
II.
E.
2­
4.
A
small
p­
value
indicates
that
logarithms
of
the
UEs
are
not
normally
distributed,
or
equivalently,
that
the
UEs
are
not
lognormally
distributed.
Both
the
granular
inhalation
UE
and
dust
inhalation
UE
datasets
resulted
in
S­
W
statistics
with
p­
values
less
than
0.05.
II.
E.
2
­
Page
4
of
46
Table
II.
E.
2­
1
Granular
Rotary
Spreader
UE
Data
(
OMA003)
Used
for
Lawn
Care
Scenario
Dermal
UE
Values
(
mg/
lb
ai)
Inhalation
UE
Values
(
mg/
lb
ai)
0.529
0.0001
0.392
0.0008
0.668
0.0011
0.692
0.0027
0.329
0.0001
0.373
0.0007
0.363
0.0007
0.595
0.0014
0.339
0.0007
0.563
0.0014
0.712
0.0026
0.253
0.0006
0.787
0.0035
0.514
0.0033
0.999
0.0015
0.412
0.0008
0.427
0.0007
0.917
0.0011
0.757
0.0010
0.827
0.0008
0.620
0.0006
0.730
0.0003
0.551
0.0006
2.104
0.0011
1.363
0.0032
0.915
0.0025
0.522
0.0007
6.980
0.0008
0.462
0.0007
1.022
0.0003
II.
E.
2
­
Page
5
of
46
Table
II.
E.
2­
2
Dust
Shaker/
Powder
UE
Data
(
MRID
44459801)
Used
for
Lawn
Care
Scenarios
Dermal
UE
Values
(
mg/
lb
ai)
Inhalation
UE
Values
(
mg/
lb
ai)
673
2.27
588
0.60
276
1.38
129
2.23
176
0.30
94
0.61
236
4.87
229
0.01
85
2.11
69
0.38
82
2.14
258
0.66
51
1.99
1388
14.27
40
0.13
280
1.09
43
1.40
36
0.57
219
2.28
59
0.26
Table
II.
E.
2­
3
Hose
End
Sprayer
on
Turf
UE
Data
(
OMA004)
Used
for
Lawn
Care
Scenarios
Dermal
UE
Values
(
mg/
lb
ai)
Inhalation
UE
Values
(
mg/
lb
ai)
0.21
0.019
6.76
0.026
32.61
0.065
1.84
0.013
3.09
0.030
3.16
0.037
1.22
0.027
1.36
0.030
1.29
0.003
0.99
0.019
9.38
0.034
1.16
0.006
3.20
0.021
9.69
0.045
5.42
0.061
12.89
0.005
1.92
0.002
8.93
0.004
3.68
0.017
11.05
0.008
0.08
0.001
23.03
0.001
4.51
0.003
0.22
0.015
2.83
0.029
1.20
0.003
8.60
0.007
0.41
0.003
23.66
0.014
0.17
0.004
II.
E.
2
­
Page
6
of
46
Table
II.
E.
2­
4
Lognormal
Distributions
of
UEs
Used
for
Lawn
Care
Scenarios
Additionally,
probability
plots
were
used
to
qualitatively
assess
the
appropriateness
of
the
lognormal
assumptions.
Generally
a
probability
plot
displays
the
actual
values
of
a
dataset
(
represented
as
points)
and
their
expected
values
(
represented
as
a
line)
for
the
specified
distribution.
The
closer
the
actual
values
are
to
their
expected
values
(
i.
e.
the
more
the
actual
values
approximate
a
straight
line),
the
more
likely
the
dataset
is
of
the
specified
distribution.
The
probability
plots
for
the
UE
datasets
are
provided
in
Figures
II.
E.
2­
1
through
6.
For
the
granular
dermal
UE
dataset,
the
probability
plot
indicates
that
the
small
S­
W
p­
value
is
due
to
one
very
high
value;
whereas
for
the
dust
inhalation
UE
dataset,
one
very
low
value
results
in
a
small
S­
W
p­
value.
The
other
datasets
are
reasonably
approximated
by
lognormal
distributions.

Figure
II.
E.
2­
1
Lognormal
Probability
Plot
of
Granular
Rotary
Spreader
Dermal
UE
Data
(
OMA003)

ln_
rot_
sprdr_
SSSP
5
6
7
8
9
10
.01
.05
.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
2
Lognormal
Probability
Plot
of
Granular
Rotary
Spreader
Inhalation
UE
Data
(
OMA003)

ln_
rot_
sprdr_
inhal
­
2
­
1.5
­
1
­
0.5
0
0.5
1
1.5
2
.01
.05.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Application
Method
Exposure
Route
Unit
Exposure
Distribution
(
mg/
lb
ai)
Shapiro­
Wilk
p­
value
Dermal
LN(
0.809,
0.570)
0.0011
Granular
Rotary
Spreader
Inhalation
LN(
0.0013,
0.0013)
0.0511
Dermal
LN(
247,
333)
0.3691
Dust
Shaker/
Powder
Inhalation
LN(
2.94,
9.54)
0.0354
Dermal
LN(
8.44,
26.2)
0.3630
Hose
End
Sprayer
(
RTU)
on
Turf
Inhalation
LN(
0.022,
0.040)
0.1890
NOTES:

LN(
µ
,
 )
represents
a
lognormal
distribution
with
mean=
µ
and
standard
deviation=
 .

For
lawn
scenarios,
information
was
derived
from
chemical­
specific
data
and
studies
conducted
by
the
ORETF
(
Outdoor
Residential
Exposure
Task).
II.
E.
2
­
Page
7
of
46
Figure
II.
E.
2­
3
Lognormal
Probability
Plot
of
Dust
Shaker/
Powder
Dermal
UE
Data
(
MRID
44459801)

ln_
dust_
garden
3
4
5
6
7
8
.01
.05.10
.25
.50
.75
.90.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
4
Lognormal
Probability
Plot
of
Dust
Shaker/
Powder
Inhalation
UE
Data
(
MRID
44459801)

ln_
dust_
gard_
inh
1
2
3
4
5
6
7
8
9
10
11
.01
.05
.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
5
Lognormal
Probability
Plot
of
HES
Sprayer
on
Turf
Dermal
UE
Data
(
OMA004)

ln_
OMA004_
SSSP
4
5
6
7
8
9
10
11
12
13
.01
.05
.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
6
Lognormal
Probability
Plot
of
Hose
End
Sprayer
on
Turf
Inhalation
UE
Data
(
OMA004)

ln_
OMA004_
inh
­
1
0
1
2
3
4
5
6
.01
.05
.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Residue
Data
Study
Summaries
MRID
451143­
01
(
Carbaryl
Turf
Transferable
Residue
Study):
A
TTR
study
was
conducted
at
individual
sites
in
three
states
using
the
ORETF
roller
sampling
method.
The
data
used
in
this
assessment
was
from
the
Georgia
site.
Bermudagrass
was
the
variety
of
turfgrass
treated
at
the
Georgia
site.
Field
work
took
place
over
three
week
intervals
at
each
site.
Applications
were
made
and
II.
E.
2
­
Page
8
of
46
samples
were
collected
essentially
in
October
of
1998
Georgia.
Two
applications
were
made
7
days
apart
at
each
site.
All
applications
in
this
study
were
completed
at
a
rate
of
8.17
lb
ai/
acre.
Applications
were
made
with
typical
groundboom
sprayers
using
approximately
55
and
31
gallons
of
water
per
acre,
respectively.
All
applications
were
made
using
Dragon
Sevin
Liquid
which
is
a
flowable
concentrate
formulation
that
contains
carbaryl
at
a
nominal
concentration
of
21
percent
by
weight
or
2
lb
ai/
gallon.

There
was
approximately
from
1
inch
up
to
2.7
inches
of
irrigation
water
on
the
day
of
the
final
application.
Additionally,
on
the
day
of
the
final
application,
rain
was
noted
that
ranged
in
accumulations
of
0.36
inches.
Mowing
events
were
not
noted
in
the
data
from
the
Georgia
site.
Triplicate
TTR
samples
were
collected
using
the
ORETF
roller
method
at
8
intervals
out
to
14
days
after
the
last
application.
All
but
two
samples
were
collected
during
the
1st
week
of
the
study.
In
all
cases,
residue
levels
exceeded
the
LOQ
at
14
days
after
application.

Statistical
Details
Turf
transferable
residues
(
TTR)
values
are
assumed
to
degrade
exponentially
over
time
(
i.
e.
degrade
by
a
constant
proportion
for
any
given
time
interval).
In
order
to
estimate
the
initial
TTR
value
(
i.
e.
TTR
value
at
day
zero)
and
the
halflife
of
the
liquid
formulation
of
carbaryl,
the
natural
logarithms
of
the
27
(
3
samples
X
9
days)
individual
TTR
samples
(
Table
II.
E.
2­
5)
from
the
Georgia
site
were
linearly
regressed
on
the
day
of
sample
collection.
The
form
of
the
linear
regression
is
given
below.

t
 
 
ln(
y)
1
0
+
=

The
linear
regression
parameters
were
then
used
to
calculate
initial
DFR
value
(
A0)
and
the
half­
life
(
T1/
2)
using
formulae
given
below.

0
0
 
A
=

1
 
ln(
2)
T
2
1
 
=
II.
E.
2
­
Page
9
of
46
Table
II.
E.
2­
5
Liquid
Formulation
TTR
Data
(
MRID
#
451143­
01,
Georgia
site)
Used
for
Lawn
Care
Scenario
Day
TTR
Values
(
mg/
cm2)

0.00130
0.00122
0
0.00152
0.00067
0.00073
0.5
0.00147
0.00041
0.00042
1
0.00047
0.00020
0.00028
2
0.00023
0.00014
0.00027
3
0.00050
0.00040
0.00023
5
0.00010
0.00011
0.00013
7
0.00031
0.00003
0.00015
10
0.00022
0.00007
0.00002
14
0.00015
Transfer
Coefficient
Data
Study
Summaries
Black,
1993:
A
study
by
Black
(
1993),
which
investigated
dermal
exposure
values
of
young
children
who
were
exposed
to
a
non­
toxic
substance,
was
used
to
represent
the
spray
application
scenario.
In
this
study,
children
performed
unscripted
activities
on
turf
treated
with
a
non­
toxic
substance
used
as
a
whitening
agent
in
fabrics.
The
subjects
of
the
study
were
14
children
aged
four
to
nine
years
old.
In
this
study,
children
were
provided
toys
and
their
activities
were
recorded
as
they
performed
unscripted
activities
for
a
period
of
one
half
hour.
Activities
recorded
were
grouped
into
the
following
classifications:


Upright
(
standing,
walking,
jumping
and
running)
II.
E.
2
­
Page
10
of
46

Sitting
(
straight­
up,
cross
legged,
kneeling,
crouching
and
crawling)


Lying
(
prone
or
supine)

Dermal
exposure
was
measured
by
fluorescent
measurement
technology
described
in
Fenske
et
al.,
(
1986).
Measurements
on
various
body
parts
were
expressed
as
ug/
body
part
(
e.
g.,
hand,
face,
etc.)
and
as
concentration
(
ug/
cm2).

Vaccaro,
1993:
In
a
second
study
(
Vaccaro,
1993)
in
which
a
liquid
formulation
was
used,
eight
adults
performed
structured
activities
intended
to
mimic
a
child's
activities
(
including
walking/
running,
sleeping,
crawling,
and
sitting
on
turf).
The
subjects
performed
these
activities
for
a
period
of
four
hours
beginning
four
hours
after
the
turf
had
dried.
Turf
had
been
treated
earlier
with
a
sprayable
form
of
chlorpyrifos
and
exposure
was
estimated
in
the
study
by
monitoring
the
amount
of
a
chlorpyrifos
metabolite
 
excreted
over
the
following
period
of
6
days.
This
method
directly
measured
internal
dose
and
was
used
to
back­
calculate
a
generic
"
to
the
skin"
transfer
coefficient
by
using
chemical
specific
dermal
absorption
data
for
chlorpyrifos
(
Nolan
et
al.,
1984).

Statistical
Details
Distributional
parameters
were
estimated
for
the
combined
children
transfer
coefficient
(
TC)
values
from
the
Black
and
Vaccaro
studies
and
the
adult
TC
values
from
the
Vaccaro
study
(
Table
II.
E.
2­
6).
All
TC
values
were
expressed
as
square
centimeters
per
hour
(
cm2/
hr).
Both
children
and
adult
TCs
were
assumed
to
be
lognormally
distributed
(
i.
e.
fitted
with
a
lognormal
distribution).
For
each
dataset,
the
shape
(
 )
and
scale
(
 )
lognormal
parameters
were
estimated
by
calculating
the
mean
and
standard
deviation
of
the
natural
logarithms
(
base
e)
of
the
TCs.
Parametric
estimates
of
the
arithmetic
mean
(
µ
)
and
standard
deviation
(
 )
of
the
lognormal
distribution
were
then
calculated
based
on
the
shape
and
scale
parameter
estimates.
The
formulae
used
to
calculate
the
mean
and
standard
deviation
are
given
below.

)
 
exp(
 
µ
2
2
1
+
=

1
)
exp(
 
µ
 
2
 
=

Shapiro­
Wilk
(
S­
W)
normality
test
statistics
were
used
to
assess
the
lognormal
assumption
implicit
in
the
parametric
calculations
of
the
mean
and
standard
deviation.
The
means,
standard
deviations,
and
p­
values
of
the
S­
W
statistics
are
provided
in
Table
II.
E.
2­
7.
A
small
p­
value
indicates
that
logarithms
of
the
TCs
are
not
normally
distributed,
or
equivalently,
that
the
TCs
are
not
lognormally
distributed.
For
both
children
and
adult
TC
datasets,
the
S­
W
p­
values
are
greater
than
0.05.
II.
E.
2
­
Page
11
of
46
Table
II.
E.
2­
6
Liquid
Formulation
TC
Data
Used
for
Lawn
Care
Scenarios
Population
Study
TC
Values
(
cm2/
hr)

1240
2507
2673
3251
3665
4164
4877
Vaccaro
4905
2844
3594
3776
4051
4103
4357
4902
6812
8395
8746
9119
9885
10713
Children
Black
16008
3348
6770
7217
8779
9895
11243
13169
Adult
Vaccaro
13243
Table
II.
E.
2­
7
Lognormal
Distributions
of
TCs
Used
for
Lawn
Care
Scenarios
Application
Method
Exposure
Route
Population
Transfer
Coefficient
Distribution
(
cm2/
hr)
Shapiro­
Wilk
p­
value
Dermal
Children
LN(
5709,
3634)
0.6127
Hose
End
Sprayer*
Dermal
Adult
LN(
9445,
4509)
0.2035
NOTES:
LN(
µ
,
 )
represents
a
lognormal
distribution
with
mean=
µ
and
standard
deviation=
 .

*
Liquid
formulation
data
was
used
for
granular
exposure
scenarios.
II.
E.
2
­
Page
12
of
46
Additionally,
probability
plots
were
used
to
qualitatively
assess
the
appropriateness
of
the
lognormal
assumptions.
Generally
a
probability
plot
displays
the
actual
values
of
a
dataset
(
represented
as
points)
and
their
expected
values
(
represented
as
a
line)
for
the
specified
distribution.
The
closer
the
actual
values
are
to
their
expected
values
(
i.
e.
the
more
the
actual
values
approximate
a
straight
line),
the
more
likely
the
dataset
is
of
the
specified
distribution.
The
probability
plots
for
the
TC
datasets
are
provided
in
Figures
II.
E.
2­
7
and
8.
The
probability
plots
indicate
the
both
children
and
adult
TC
datasets
are
reasonably
approximated
by
lognormal
distributions.

Figure
II.
E.
2­
7
Lognormal
Probability
Plot
of
Liquid
Formulation
Child
TC
Data
liq_
lawn_
TC_
child
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
.01
.05.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
8
Lognormal
Probability
Plot
of
Liquid
Formulation
Adult
TC
Data
ln_
liq_
lawn_
adult
8
8.5
9
9.5
10
.01
.05
.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
2.
Vegetable
Garden
Exposure
Scenarios
Unit
Exposure
Data
Study
Summaries
MRID
444598­
01
(
Carbaryl
Applications
To
Vegetables
Gardens):
The
data
collected
reflect
the
dermal
and
respiratory
exposure
of
homeowners
mixing,
loading
and
applying
RP­
2
Liquid
(
21%),
a
Carbaryl
end­
use
product.
Applications
were
made
by
volunteers
to
two
18
foot
rows
of
tomatoes
and
one
18
foot
row
of
cucumber.
The
only
test
field
was
located
in
Florida.
For
this
study,
RP­
2
Liquid
(
21%)
exposures
were
monitored
using
hose­
end
sprayers
and
low­
pressure
hand
wand
sprayers.
Exposures
to
Sevin
®
10
Dust,
using
a
separate
duster
device
that
required
transfer
from
the
package
and
Sevin
®
Ready
To
Use
Insect
Spray
(
RTU)
in
a
trigger
sprayer
package
were
also
monitored.
Exposure
for
each
spray
method/
product
combination
was
monitored
using
40
handlers
(
replicates).
Of
the
40
replicates
per
spray
method/
product
combination,
20
wore
household
latex
gloves
and
20
performed
tasks
without
II.
E.
2
­
Page
13
of
46
gloves.
The
20
dust
product
replicates
loaded
the
dusters
and
applied
without
gloves
only.

Each
replicate
opened
the
end­
use
product,
added
it
to
the
application
implement
(
except
the
RTU
product),
adjusted
the
setting
and
applied
it
to
the
vegetable
rows.
After
application
to
the
vegetable
rows,
dosimeters
were
collected.
Inhalation
exposure
was
monitored
with
personal
air
sampling
pumps
with
OVS
tubes
attached
to
the
shirt
collar
in
the
breathing
zone.
Dermal
exposure
was
assessed
by
extraction
of
carbaryl
from
inner
and
outer
100
percent
cotton
dosimeters,
face/
neck
wipes,
and
glove
and
hand
washes.
The
inner
and
outer
dosimeters
were
segmented
into:
lower
and
upper
arms,
lower
and
upper
legs,
front
and
back
torso.

Dermal
exposure
was
determined
by
adding
the
values
from
the
bare
hand
rinses,
face/
neck
wipes
to
the
outer
dosimeter
lower
legs
and
lower
arms
plus
the
inner
dosimeter
front
and
rear
torso,
upper
legs,
lower
legs,
lower
arms,
and
upper
arms.
This
accounts
for
the
residential
handlers
with
barehands
wearing
short­
sleeved
shirt
and
short
pants.

Statistical
Details
Distributional
parameters
were
estimated
for
the
dermal
and
inhalation
unit
exposure
(
UE)
values
for
dust
(
Table
II.
E.
2­
8),
trigger
pump
sprayer
(
Table
II.
E.
2­
9),
and
liquid
hose­
end
sprayer
(
Table
II.
E.
2­
10)
applications
of
Carbaryl.
All
dermal
and
inhalation
UE
values
represent
milligrams
exposure
per
pound
of
active
ingredient
of
a
pesticide
handled.
All
UEs
were
assumed
to
be
lognormally
distributed
(
i.
e.
fitted
with
a
lognormal
distribution).
For
each
dataset,
the
shape
(
 )
and
scale
(
 )
lognormal
parameters
were
estimated
by
calculating
the
mean
and
standard
deviation
of
the
natural
logarithms
(
base
e)
of
the
UEs.
Parametric
estimates
of
the
arithmetic
mean
(
µ
)
and
standard
deviation
(
 )
of
the
lognormal
distribution
were
then
calculated
based
on
the
shape
and
scale
parameter
estimates.
The
formulae
used
to
calculate
the
mean
and
standard
deviation
are
given
below.

)
 
exp(
 
µ
2
2
1
+
=

1
)
exp(
 
µ
 
2
 
=

Shapiro­
Wilk
(
S­
W)
normality
test
statistics
were
used
to
assess
the
lognormal
assumption
implicit
in
the
parametric
calculations
of
the
mean
and
standard
deviation.
The
means,
standard
deviations,
and
p­
values
of
the
S­
W
statistics
are
provided
in
Table
II.
E.
2­
11.
A
small
p­
value
indicates
that
logarithms
of
the
UEs
are
not
normally
distributed,
or
equivalently,
that
the
UEs
are
not
lognormally
distributed.
Both
the
dust
inhalation
UE
and
trigger
pump
inhalation
UE
datasets
resulted
in
S­
W
statistics
with
p­
values
less
than
0.05.
II.
E.
2
­
Page
14
of
46
Table
II.
E.
2­
8
Dust
Shaker/
Powder
UE
Data
(
MRID
44459801)
Used
for
Vegetable
Garden
Scenarios
Dermal
UE
Values
(
mg/
lb
ai)
Inhalation
UE
Values
(
mg/
lb
ai)
673
2.27
588
0.60
276
1.38
129
2.23
176
0.30
94
0.61
236
4.87
229
0.01
85
2.11
69
0.38
82
2.14
258
0.66
51
1.99
1388
14.27
40
0.13
280
1.09
43
1.40
36
0.57
219
2.28
59
0.26
Table
II.
E.
2­
9
Trigger
Pump
Sprayer
UE
Data
(
MRID
44459801)
Used
for
Vegetable
Garden
Scenarios
Dermal
UE
Values
(
mg/
lb
ai)
Inhalation
UE
Values
(
mg/
lb
ai)
129
0.275
59
0.255
250
0.104
132
0.168
145
0.180
91
0.032
165
0.180
77
0.200
24
0.033
50
0.032
24
0.033
100
0.086
23
0.032
24
0.110
20
0.035
218
0.032
9
0.032
18
0.032
41
0.032
23
0.032
II.
E.
2
­
Page
15
of
46
Table
II.
E.
2­
10
Hose
End
Sprayer
UE
Data
(
MRID44459801)
for
Vegetable
Garden
Scenarios
Dermal
UE
Values
(
mg/
lb
ai)
Inhalation
UE
Values
(
mg/
lb
ai)
31
0.0022
47
0.0009
21
0.0016
77
0.0028
58
0.0014
76
0.0030
25
0.0032
31
0.0044
19
0.0017
17
0.0013
33
0.0010
84
0.0041
24
0.0023
56
0.0009
8
0.0027
199
0.0044
163
0.0014
11
0.0007
21
0.0044
7
0.0028
Table
II.
E.
2­
11
Lognormal
Distributions
of
UEs
Used
for
Vegetable
Garden
Scenarios
Additionally,
probability
plots
were
used
to
qualitatively
assess
the
appropriateness
of
the
lognormal
assumptions.
Generally
a
probability
plot
displays
the
actual
values
of
a
dataset
(
represented
as
points)
and
their
expected
values
(
represented
as
a
line)
for
the
specified
distribution.
The
closer
the
actual
Application
Method
Exposure
Route
Unit
Exposure
Distribution
(
mg/
lb
ai)
Shapiro­
Wilk
p­
value
Dermal
LN(
247,
333)
0.3691
Dust
Shaker/
Powder
Inhalation
LN(
2.94,
9.54)
0.0354
Dermal
LN(
86,
107)
0.2191
Trigger
Pump
Sprayer
Inhalation
LN(
0.104,
0.137)*
0.0003
Dermal
LN(
51,
58)
0.8266
Hose­
End
Sprayer
Inhalation
LN(
0.0024,
0.0015)
0.2075
NOTES:
LN(
µ
,
 )
represents
a
lognormal
distribution
with
mean=
µ
and
standard
deviation=
 .
*
The
mean
and
standard
deviation
represent
MLE­
based
estimates.
II.
E.
2
­
Page
16
of
46
values
are
to
their
expected
values
(
i.
e.
the
more
the
actual
values
approximate
a
straight
line),
the
more
likely
the
dataset
is
of
the
specified
distribution.
The
probability
plots
for
the
UE
datasets
are
provided
in
Figures
II.
E9
through
14.
For
the
dust
inhalation
UE
dataset,
the
probability
plot
indicates
that
the
small
SW
p­
value
is
due
to
one
very
low
value;
whereas
for
the
trigger
pump
inhalation
UE
dataset,
several
low
values
account
for
the
small
S­
W
p­
value.
The
other
datasets
are
reasonably
approximated
by
lognormal
distributions.

F
igure
II.
E.
2­
9
Lognormal
Probability
Plot
of
Dust
Shaker/
Powder
Dermal
UE
Data
(
MRID
44459801)

ln_
dust_
garden
3
4
5
6
7
8
.01
.05.10
.25
.50
.75
.90.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
10
Lognormal
Probability
Plot
of
Dust
Shaker/
Powder
Inhalation
UE
Data
(
MRID
44459801)

ln_
dust_
gard_
inh
1
2
3
4
5
6
7
8
9
10
11
.01
.05
.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
11
Lognormal
Probability
Plot
of
Trigger
Pump
Sprayer
UE
Dermal
Data
(
MRID
44459801)

ln_
RTU_
garden
1
2
3
4
5
6
7
.01
.05.10
.25
.50
.75
.90.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
12
Lognormal
Probability
Plot
of
Trigger
Pump
Sprayer
UE
Inhalation
Data
(
MRID
44459801)

ln_
RTU_
gard_
inh
3
4
5
6
.01
.05.10
.25
.50
.75
.90.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
II.
E.
2
­
Page
17
of
46
Figure
II.
E.
2­
13
Lognormal
Probability
Plot
of
Hose
End
Sprayer
Dermal
UE
Data
(
MRID44459801)

ln_
hose_
garden
1
2
3
4
5
6
.01
.05.10
.25
.50
.75
.90.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
14
Lognormal
Probability
Plot
of
Hose
End
Sprayer
Inhalation
UE
Data
(
MRID44459801)

ln_
hose_
gard_
inh
­
1
­
0.5
0
0.5
1
1.5
2
.01
.05.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
For
the
trigger
pump
inhalation
UE
dataset,
11
out
of
20
samples
were
reported
as
approximately
the
same
value.
All
11
samples
from
the
inhalation
monitors
were
reported
as
0.07
µ
g
with
slightly
different
amounts
of
active
ingredient
handled
by
the
study
subjects,
which
resulted
in
slightly
different
UE
(
mg/
lb
ai)
values.
The
value
0.07
µ
g
was
assumed
to
be
half
the
LOQ.
The
mean
and
standard
deviation
estimated
for
the
trigger
pump
inhalation
UE
dataset
are
based
on
maximum
likelihood
estimation
(
MLE)
procedures
assuming
the
dataset
represents
a
sample
from
a
censored
lognormal
distribution.

Residue
Data
Study
Summaries
MRID
450059­
09
(
Carbaryl
Sunflower
DFR
Study):
The
field
phase
of
this
study
was
conducted
at
a
single
site
near
Northwood,
North
Dakota.
The
field
phase
of
the
study
was
conducted
during
the
period
from
July
20
to
August
25,
1998.
Sample
analyses
were
completed
by
December
1998.
A
fixed­
wing
aircraft
was
used
to
make
2
applications
of
Sevin
XLR
Plus,
a
liquid
flowable
formulation,
7
days
apart
at
an
application
rate
of
1.5
lb
ai/
acre.
Spray
volume
was
3
gallons
of
water
per
acre.
The
sunflower
plants
were
approximately
4
feet
tall
and
were
spaced
approximately
0.5
feet
within
each
row
while
the
rows
were
spaced
2.5
feet
apart
(
i.
e.,
~
35000
plants/
acre).
No
significant
precipitation
was
observed
in
this
study
until
at
least
14
days
after
application.

DFR
samples
were
collected
out
to
28
days
after
the
last
application
using
the
Iwata
method
(
i.
e.,
a
total
surface
area
sampled
of
400
cm2/
sample
collected
with
a
1
inch
diameter
Birkestrand
leaf
punch
and
dislodged
with
a
0.01
percent
Aerosol
solution).
There
were
still
measurable
residues
28
days
after
II.
E.
2
­
Page
18
of
46
application.
The
percent
transferability
of
the
0
day
sample
was
32
percent
of
the
application
rate.

Statistical
Details
Dislodgeable
foliar
residue
(
DFR)
values
are
assumed
to
degrade
exponentially
over
time
(
i.
e.
degrade
by
a
constant
proportion
for
any
given
time
interval).
In
order
to
estimate
the
initial
DFR
value
(
i.
e.
DFR
value
at
day
zero)
and
the
halflife
of
the
liquid
formulation
of
Carbaryl,
the
natural
logarithms
of
the
30
(
3
samples
X
10
days)
individual
DFR
samples
(
Table
II.
E.
2­
12)
from
the
sunflower
study
were
linearly
regressed
on
the
day
of
sample
collection.
The
form
of
the
linear
regression
is
given
below.

t
 
 
ln(
y)
1
0
+
=

The
linear
regression
parameters
were
then
used
to
calculate
initial
DFR
value
(
A0)
and
the
half­
life
(
T1/
2)
using
formulae
given
below.

0
0
 
A
=

1
 
ln(
2)
T
2
1
 
=
II.
E.
2
­
Page
19
of
46
Table
II.
E.
2­
12
Liquid
Formulation
DFR
Data
(
MRID
45005909)
Used
for
Vegetable
Garden
Scenarios
Day
DFR
Values
(
mg/
cm2)

0.00503
0.00615
0
0.00488
0.00425
0.00515
1
0.00508
0.00415
0.00295
2
0.00380
0.00393
0.00330
3
0.00483
0.00498
0.00463
4
0.00418
0.00241
0.00205
5
0.00310
0.00308
0.00308
6
0.00320
0.00283
0.00213
7
0.00288
0.00139
0.00116
14
0.00108
0.00020
0.00010
28
0.00005
Transfer
Coefficient
Data
Study
Summaries
MRID
45344501
(
Chrysanthemum
Pinching):
This
study
was
conducted
with
volunteer
workers
pinching
buds
from
greenhouse
chrysanthemums
after
two
treatments
with
the
active
ingredient
(
ai)
diazainon,
formulated
as
an
emulsifiable
concentrate
called
Diazinon
AG600
WBC
®
.
Dermal
and
inhalation
data
were
collected,
together
with
concurrent
dislodgeable
foliar
residue
(
DFR)
data.
Potential
exposures
were
measured
using
whole­
body
dosimeters
(
outer
and
inner
dosimetry),
hand
washes,
and
face/
neck
wipes
for
dermal
exposure
and
personal
sampling
pumps
for
inhalation
exposure.
Transfer
coefficients
were
calculated
for
potential
and
total
dermal
exposure.
II.
E.
2
­
Page
20
of
46
MRID
451917­
01
(
Cabbage
Weeding
Study):
This
study
was
conducted
with
volunteer
workers
weeding
commercially
grown
cabbage
after
two
treatments
with
the
active
ingredient
(
ai)
carbaryl,
formulated
as
the
emulsifiable
concentrate
Sevin
®
XLR
Plus.
The
potential
dermal
and
respiratory
exposure
during
reentry
was
assessed
at
an
established
commercial
cabbage
field
in
San
Joaquin
Valley
of
California
by
using
whole­
body
dosimetry,
hand
washes,
face/
neck
wipes,
and
a
personal
air
sampling
pump.
Transfer
coefficients
for
potential
and
total
dermal
exposure
were
calculated.

3.
Ornamental
Plants
and
Shrubs
Exposure
Scenarios
Unit
Exposure
Data
Study
Summaries
MRID
444598­
01
(
Carbaryl
Applications
To
Vegetable
Gardens):
The
data
collected
reflect
the
dermal
and
respiratory
exposure
of
homeowners
mixing,
loading
and
applying
RP­
2
Liquid
(
21%),
a
Carbaryl
end­
use
product.
Applications
were
made
by
volunteers
to
two
18
foot
rows
of
tomatoes
and
one
18
foot
row
of
cucumber.
The
only
test
field
was
located
in
Florida.
For
this
study,
RP­
2
Liquid
(
21%)
exposures
were
monitored
using
hose­
end
sprayers
and
low­
pressure
hand
wand
sprayers.
Exposures
to
Sevin
®
10
Dust,
using
a
separate
duster
device
that
required
transfer
from
the
package
and
Sevin
®
Ready
To
Use
Insect
Spray
(
RTU)
in
a
trigger
sprayer
package
were
also
monitored.
Exposure
for
each
spray
method/
product
combination
was
monitored
using
40
handlers
(
replicates).
Of
the
40
replicates
per
spray
method/
product
combination,
20
wore
household
latex
gloves
and
20
performed
tasks
without
gloves.
The
20
dust
product
replicates
loaded
the
dusters
and
applied
without
gloves
only.

Each
replicate
opened
the
end­
use
product,
added
it
to
the
application
implement
(
except
the
RTU
product),
adjusted
the
setting
and
applied
it
to
the
vegetable
rows.
After
application
to
the
vegetable
rows,
dosimeters
were
collected.
Inhalation
exposure
was
monitored
with
personal
air
sampling
pumps
with
OVS
tubes
attached
to
the
shirt
collar
in
the
breathing
zone.
Dermal
exposure
was
assessed
by
extraction
of
carbaryl
from
inner
and
outer
100
percent
cotton
dosimeters,
face/
neck
wipes,
and
glove
and
hand
washes.
The
inner
and
outer
dosimeters
were
segmented
into:
lower
and
upper
arms,
lower
and
upper
legs,
front
and
back
torso.

Dermal
exposure
was
determined
by
adding
the
values
from
the
bare
hand
rinses,
face/
neck
wipes
to
the
outer
dosimeter
lower
legs
and
lower
arms
plus
the
inner
dosimeter
front
and
rear
torso,
upper
legs,
lower
legs,
lower
arms,
and
upper
arms.
This
accounts
for
the
residential
handlers
with
barehands
wearing
short­
sleeved
shirt
and
short
pants.
II.
E.
2
­
Page
21
of
46
MRID
445185­
01
(
Carbaryl
Applications
To
Trees
And
Shrubs
Study):
Applications
of
Sevin
Liquid
®
Carbaryl
insecticide
[
RP­
2
liquid
(
21%)]
were
made
by
volunteers
to
two
young
citrus
trees
and
two
shrubs
in
each
replicate
that
was
monitored
in
the
study.
The
test
field
was
located
only
in
Florida.
Twenty
(
20)
replicates
were
monitored
using
hose­
end
sprayer
(
Ortho
®
DIAL
or
Spray
®
hose
end
sprayer),
and
20
replicates
were
monitored
using
hand
held
pump
sprayers
(
low­
pressure
hand
wands).

Each
replicate
opened
the
end­
use
product,
added
it
to
the
hose­
end
sprayer
or
hand
held
pump
and
then
applied
it
to
the
trees
and
shrubs.
After
application
to
two
trees
and
two
shrubs
dosimeters
were
collected.
Inhalation
exposure
was
monitored
with
personal
air
sampling
pumps
with
OVS
tubes
attached
to
the
shirt
collar
in
the
breathing
zone.
Dermal
exposure
was
assessed
by
extraction
of
Carbaryl
from
inner
and
outer
100
percent
cotton
dosimeters.
The
inner
and
outer
dosimeters
were
segmented
into:
lower
and
upper
arms,
lower
and
upper
legs,
front
and
back
torso.
No
gloves
were
worn
therefore
hand
exposure
was
assessed
with
400
ml
handwash
with
0.01
percent
Aerosol
OT­
75
sodium
dioctyl
sulfosuccinate
(
OTS).
One
hundred
(
100)
percent
cotton
handkerchiefs
wetted
with
25
ml
OTS
were
used
to
wipe
face
and
neck
to
determine
exposure.

The
dermal
exposure
was
calculated
by
adding
the
values
from
the
hand
rinses,
face/
neck
wipes
to
the
outer
dosimeter
lower
legs
and
lower
arms
plus
the
inner
dosimeter
front
and
rear
torso,
upper
legs,
lower
legs,
lower
arms,
and
upper
arms.
This
accounts
for
the
residential
handlers
with
barehands
wearing
shortsleeved
shirt
and
short
pants.

Statistical
Details
Distributional
parameters
were
estimated
for
the
dermal
and
inhalation
unit
exposure
(
UE)
values
for
dust
(
Table
II.
E.
2­
13),
trigger
pump
sprayer
(
Table
II.
E.
2­
14),
and
liquid
hand
wand
sprayer
(
Table
II.
E.
2­
15)
applications
of
carbaryl.
All
dermal
and
inhalation
UE
values
represent
milligrams
exposure
per
pound
of
active
ingredient
of
a
pesticide
handled.
All
UEs
were
assumed
to
be
lognormally
distributed
(
i.
e.
fitted
with
a
lognormal
distribution).
For
each
dataset,
the
shape
(
 )
and
scale
(
 )
lognormal
parameters
were
estimated
by
calculating
the
mean
and
standard
deviation
of
the
natural
logarithms
(
base
e)
of
the
UEs.
Parametric
estimates
of
the
arithmetic
mean
(
µ
)
and
standard
deviation
(
 )
of
the
lognormal
distribution
were
then
calculated
based
on
the
shape
and
scale
parameter
estimates.
The
formulae
used
to
calculate
the
mean
and
standard
deviation
are
given
below.

)
 
exp(
 
µ
2
2
1
+
=

1
)
exp(
 
µ
 
2
 
=
II.
E.
2
­
Page
22
of
46
Shapiro­
Wilk
(
S­
W)
normality
test
statistics
were
used
to
assess
the
lognormal
assumption
implicit
in
the
parametric
calculations
of
the
mean
and
standard
deviation.
The
means,
standard
deviations,
and
p­
values
of
the
S­
W
statistics
are
provided
in
Table
II.
E.
2­
16.
A
small
p­
value
indicates
that
logarithms
of
the
UEs
are
not
normally
distributed,
or
equivalently,
that
the
UEs
are
not
lognormally
distributed.
The
dust
inhalation
UE,
trigger
pump
inhalation
UE,
and
hand
wand
UE
datasets
resulted
in
S­
W
statistics
with
p­
values
less
than
0.05.

Table
II.
E.
2­
13
Dust
Shaker/
Powder
UE
Data
(
MRID
44459801)
Used
for
Ornamental
Plants
and
Shrubs
Scenarios
Dermal
UE
Values
(
mg/
lb
ai)
Inhalation
UE
Values
(
mg/
lb
ai)
673
2.27
588
0.60
276
1.38
129
2.23
176
0.30
94
0.61
236
4.87
229
0.01
85
2.11
69
0.38
82
2.14
258
0.66
51
1.99
1388
14.27
40
0.13
280
1.09
43
1.40
36
0.57
219
2.28
59
0.26
Table
II.
E.
2­
14
Trigger
Pump
Sprayer
UE
Data
(
MRID
#
44459801)
Used
for
Ornamental
Plants
and
Shrubs
Scenarios
Dermal
UE
Values
(
mg/
lb
ai)
Inhalation
UE
Values
(
mg/
lb
ai)
129
0.275
59
0.255
250
0.104
132
0.168
145
0.180
91
0.032
165
0.180
77
0.200
24
0.033
50
0.032
24
0.033
100
0.086
23
0.032
24
0.110
20
0.035
218
0.032
9
0.032
18
0.032
41
0.032
23
0.032
II.
E.
2
­
Page
23
of
46
Table
II.
E.
2­
15
Hand
Wand
Sprayer
UE
Data
(
MRID
44518501)
Used
for
Ornamental
Plants
and
Shrubs
Scenarios
Dermal
UE
Values
(
mg/
lb
ai)
Inhalation
UE
Values
(
mg/
lb
ai)
25
0.004
52
0.005
129
0.004
27
0.004
348
0.005
56
0.005
118
0.004
176
0.016
44
0.009
41
0.016
46
0.004
15
0.004
36
0.004
83
0.004
78
0.025
78
0.012
46
0.004
36
0.022
25
0.004
63
0.018
Table
II.
E.
2­
16
Lognormal
Distributions
of
UEs
Used
for
Ornamental
Plants
and
Shrubs
Scenarios
Additionally,
probability
plots
were
used
to
qualitatively
assess
the
appropriateness
of
the
lognormal
assumptions.
Generally
a
probability
plot
displays
the
actual
values
of
a
dataset
(
represented
as
points)
and
their
expected
values
(
represented
as
a
line)
for
the
specified
distribution.
The
closer
the
actual
Application
Method
Exposure
Route
Unit
Exposure
Distribution
(
mg/
lb
ai)
Shapiro­
Wilk
p­
value
Dermal
LN(
247,
333)
0.3691
Dust
Shaker/
Powder
Inhalation
LN(
2.94,
9.54)
0.0354
Dermal
LN(
86,
107)
0.2191
Trigger
Pump
Sprayer
Inhalation
LN(
0.104,
0.137)*
0.0003
Dermal
LN(
74,
64)
0.7478
Hand
Wand
Sprayer
Inhalation
LN(
0.0089,
0.0102)*
0.0005
NOTES:
LN(
µ
,
 )
represents
a
lognormal
distribution
with
mean=
µ
and
standard
deviation=
 .
*
The
mean
and
standard
deviation
represent
MLE­
based
estimates.
II.
E.
2
­
Page
24
of
46
values
are
to
their
expected
values
(
i.
e.
the
more
the
actual
values
approximate
a
straight
line),
the
more
likely
the
dataset
is
of
the
specified
distribution.
The
probability
plots
for
the
UE
datasets
are
provided
in
Figures
II.
E.
2­
15
through
20.
For
the
dust
inhalation
UE
dataset,
the
probability
plot
indicates
that
the
small
SW
p­
value
is
due
to
one
very
low
value;
whereas
for
the
trigger
pump
inhalation
and
hand
wand
inhalation
UE
datasets,
several
low
values
account
for
the
small
S­
W
p­
values.
The
other
datasets
are
reasonably
approximated
by
lognormal
distributions.

Figure
II.
E.
2­
15
Lognormal
Probability
Plot
of
Dust
Shaker/
Powder
Dermal
UE
Data
(
MRID
44459801)

ln_
dust_
garden
3
4
5
6
7
8
.01
.05.10
.25
.50
.75
.90.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
16
Lognormal
Probability
Plot
of
Dust
Shaker/
Powder
Inhalation
UE
Data
(
MRID
44459801)

ln_
dust_
gard_
inh
1
2
3
4
5
6
7
8
9
10
11
.01
.05
.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
17
Lognormal
Probability
Plot
of
Trigger
Pump
Sprayer
Dermal
UE
Data
(
MRID
#
44459801)

ln_
RTU_
garden
1
2
3
4
5
6
7
.01
.05.10
.25
.50
.75
.90.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
18
Lognormal
Probability
Plot
of
Trigger
Pump
Sprayer
Inhalation
UE
Data
(
MRID
#
44459801)

ln_
RTU_
gard_
inh
3
4
5
6
.01
.05
.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
II.
E.
2
­
Page
25
of
46
Figure
II.
E.
2­
19
Lognormal
Probability
Plot
of
Hand
Wand
Sprayer
Dermal
UE
Data
(
MRID
44518501)

ln_
hw_
trees
2
3
4
5
6
7
.01
.05.10
.25
.50
.75
.90.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
20
Lognormal
Probability
Plot
of
Hand
Wand
Sprayer
Inhalation
UE
Data
(
MRID
44518501)

ln_
hw_
trees_
inh
1
2
3
4
.01
.05.10
.25
.50
.75
.90.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
For
the
trigger
pump
inhalation
and
hand
wand
inhalation
UE
datasets,
11
and
13
(
respectively)
out
of
20
samples
were
reported
as
approximately
the
same
value.
All
24
samples
from
the
inhalation
monitors
were
reported
as
0.07
µ
g
with
slightly
different
amounts
of
active
ingredient
handled
by
the
study
subjects,
which
resulted
in
slightly
different
UE
(
mg/
lb
ai)
values.
The
value
0.07
µ
g
was
assumed
to
be
half
the
LOQ.
The
means
and
standard
deviations
estimated
for
the
trigger
pump
inhalation
UE
and
hand
wand
inhalation
UE
datasets
are
based
on
maximum
likelihood
estimation
(
MLE)
procedures
assuming
the
datasets
represent
samples
from
censored
lognormal
distributions.

Transfer
Coefficient
Data
Study
Summaries
MRID
45344501
(
Chrysanthemum
Pinching):
This
study
was
conducted
with
volunteer
workers
pinching
buds
from
greenhouse
chrysanthemums
after
two
treatments
with
the
surrogate
active
ingredient
(
ai)
diazainon,
formulated
as
an
emulsifiable
concentrate
called
Diazinon
AG600
WBC
®
.
Dermal
and
inhalation
data
were
collected,
together
with
concurrent
dislodgeable
foliar
residue
(
DFR)
data.
Potential
exposures
were
measured
using
whole­
body
dosimeters
(
outer
and
inner
dosimetry),
hand
washes,
and
face/
neck
wipes
for
dermal
exposure
and
personal
sampling
pumps
for
inhalation
exposure.
Transfer
coefficients
were
calculated
for
potential
and
total
dermal
exposure.

MRID
454695­
01
(
Pruning
in
Nursery
Stock):
This
study
was
conducted
with
volunteer
workers
pruning
in
a
citrus
nursery
stock
after
one
treatment
with
the
surrogate
active
ingredient
(
ai)
malathion,
formulated
as
the
emulsifiable
concentrate.
The
potential
dermal
and
respiratory
exposure
during
reentry
was
II.
E.
2
­
Page
26
of
46
assessed
at
a
citrus
nursery
in
Arizona
by
using
whole­
body
dosimetry,
hand
washes,
face/
neck
wipes,
and
a
personal
air
sampling
pump.
Dermal
and
inhalation
data
were
collected,
together
with
concurrent
dislodgeable
foliar
residue
(
DFR)
data.
Transfer
coefficients
were
calculated
for
both
potential
and
total
dermal
exposure.

Residue
Data
Study
Summaries
MRID
450059­
09
(
Carbaryl
Sunflower
DFR
Study):
The
field
phase
of
this
study
was
conducted
at
a
single
site
near
Northwood,
North
Dakota.
The
field
phase
of
the
study
was
conducted
during
the
period
from
July
20
to
August
25,
1998.
Sample
analyses
were
completed
by
December
1998.
A
fixed­
wing
aircraft
was
used
to
make
2
applications
of
Sevin
XLR
Plus,
a
liquid
flowable
formulation,
7
days
apart
at
an
application
rate
of
1.5
lb
ai/
acre.
Spray
volume
was
3
gallons
of
water
per
acre.
The
sunflower
plants
were
approximately
4
feet
tall
and
were
spaced
approximately
0.5
feet
within
each
row
while
the
rows
were
spaced
2.5
feet
apart
(
i.
e.,
~
35000
plants/
acre).
No
significant
precipitation
was
observed
in
this
study
until
at
least
14
days
after
application.

DFR
samples
were
collected
out
to
28
days
after
the
last
application
using
the
Iwata
method
(
i.
e.,
a
total
surface
area
sampled
of
400
cm2/
sample
collected
with
a
1
inch
diameter
Birkestrand
leaf
punch
and
dislodged
with
a
0.01
percent
Aerosol
solution).
There
were
still
measurable
residues
28
days
after
application.
The
percent
transferability
of
the
0
day
sample
was
32
percent
of
the
application
rate.

Statistical
Details
Dislodgeable
foliar
residue
(
DFR)
values
are
assumed
to
degrade
exponentially
over
time
(
i.
e.
degrade
by
a
constant
proportion
for
any
given
time
interval).
In
order
to
estimate
the
initial
DFR
value
(
i.
e.
DFR
value
at
day
zero)
and
the
halflife
of
the
liquid
formulation
of
carbaryl,
the
natural
logarithms
of
the
30
(
3
samples
X
10
days)
individual
DFR
samples
(
Table
II.
E.
2­
17)
from
the
sunflower
study
were
linearly
regressed
on
the
day
of
sample
collection.
The
form
of
the
linear
regression
is
given
below.

t
 
 
ln(
y)
1
0
+
=

The
linear
regression
parameters
were
then
used
to
calculate
initial
DFR
value
(
A0)
and
the
half­
life
(
T1/
2)
using
formulae
given
below.

0
0
 
A
=

1
 
ln(
2)
T
2
1
 
=
II.
E.
2
­
Page
27
of
46
Table
II.
E.
2­
17
Liquid
Formulation
DFR
Data
(
MRID
#
45005909)
Used
for
Ornamental
Plants
and
Shrubs
Scenarios
Day
DFR
Values
(
mg/
cm2)

0.00503
0.00615
0
0.00488
0.00425
0.00515
1
0.00508
0.00415
0.00295
2
0.00380
0.00393
0.00330
3
0.00483
0.00498
0.00463
4
0.00418
0.00241
0.00205
5
0.00310
0.00308
0.00308
6
0.00320
0.00283
0.00213
7
0.00288
0.00139
0.00116
14
0.00108
0.00020
0.00010
28
0.00005
4.
Fruit
Tree
Exposure
Scenarios
Unit
Exposure
Data
Study
Summaries
MRID
445185­
01
(
Carbaryl
Applications
To
Trees
And
Shrubs):
Applications
of
Sevin
Liquid
®
Carbaryl
insecticide
[
RP­
2
liquid
(
21%)]
were
made
by
volunteers
to
two
young
citrus
trees
and
two
shrubs
in
each
replicate
that
was
monitored
in
the
study.
The
test
field
was
located
only
in
Florida.
Twenty
(
20)
II.
E.
2
­
Page
28
of
46
replicates
were
monitored
using
hose­
end
sprayer
(
Ortho
®
DIAL
or
Spray
®
hose
end
sprayer),
and
20
replicates
were
monitored
using
hand
held
pump
sprayers
(
low
pressure
hand
wands).

Each
replicate
opened
the
end­
use
product,
added
it
to
the
hose­
end
sprayer
or
hand
held
pump
and
then
applied
it
to
the
trees
and
shrubs.
After
application
to
two
trees
and
two
shrubs
dosimeters
were
collected.
Inhalation
exposure
was
monitored
with
personal
air
sampling
pumps
with
OVS
tubes
attached
to
the
shirt
collar
in
the
breathing
zone.
Dermal
exposure
was
assessed
by
extraction
of
Carbaryl
from
inner
and
outer
100
percent
cotton
dosimeters.
The
inner
and
outer
dosimeters
were
segmented
into:
lower
and
upper
arms,
lower
and
upper
legs,
front
and
back
torso.
No
gloves
were
worn
therefore
hand
exposure
was
assessed
with
400
ml
handwash
with
0.01
percent
Aerosol
OT­
75
sodium
dioctyl
sulfosuccinate
(
OTS).
One
hundred
(
100)
percent
cotton
handkerchiefs
wetted
with
25
ml
OTS
were
used
to
wipe
face
and
neck
to
determine
exposure.

The
dermal
exposure
was
calculated
by
adding
the
values
from
the
hand
rinses,
face/
neck
wipes
to
the
outer
dosimeter
lower
legs
and
lower
arms
plus
the
inner
dosimeter
front
and
rear
torso,
upper
legs,
lower
legs,
lower
arms,
and
upper
arms.
This
accounts
for
the
residential
handlers
with
barehands
wearing
shortsleeved
shirt
and
short
pants.

Statistical
Details
Distributional
parameters
were
estimated
for
the
dermal
and
inhalation
unit
exposure
(
UE)
values
for
liquid
hand
wand
sprayer
(
Table
II.
E.
2­
18)
applications
of
Carbaryl.
Dermal
and
inhalation
UE
values
represent
milligrams
exposure
per
pound
of
active
ingredient
of
a
pesticide
handled.
All
UEs
were
assumed
to
be
lognormally
distributed
(
i.
e.
fitted
with
a
lognormal
distribution).
For
each
dataset,
the
shape
(
 )
and
scale
(
 )
lognormal
parameters
were
estimated
by
calculating
the
mean
and
standard
deviation
of
the
natural
logarithms
(
base
e)
of
the
UEs.
Parametric
estimates
of
the
arithmetic
mean
(
µ
)
and
standard
deviation
(
 )
of
the
lognormal
distribution
were
then
calculated
based
on
the
shape
and
scale
parameter
estimates.
The
formulae
used
to
calculate
the
mean
and
standard
deviation
are
given
below.

)
 
exp(
 
µ
2
2
1
+
=

1
)
exp(
 
µ
 
2
 
=

Shapiro­
Wilk
(
S­
W)
normality
test
statistics
were
used
to
assess
the
lognormal
assumption
implicit
in
the
parametric
calculations
of
the
mean
and
standard
deviation.
The
means,
standard
deviations,
and
p­
values
of
the
S­
W
statistics
are
provided
in
Table
II.
E.
2­
19.
A
small
p­
value
indicates
that
logarithms
of
the
UEs
are
not
normally
distributed,
or
equivalently,
that
the
UEs
are
not
lognormally
distributed.
The
hand
wand
UE
dataset
resulted
in
an
S­
W
statistic
with
a
p­
value
less
than
0.05.
II.
E.
2
­
Page
29
of
46
Table
II.
E.
2­
18
Hand
Wand
Sprayer
UE
Data
(
MRID
#
44518501)
Used
for
Fruit
Tree
Scenarios
Table
II.
E.
2­
19
Lognormal
Distributions
of
UEs
Used
for
Fruit
Tree
Scenarios
Additionally,
probability
plots
were
used
to
qualitatively
assess
the
appropriateness
of
the
lognormal
assumptions.
Generally
a
probability
plot
displays
the
actual
values
of
a
dataset
(
represented
as
points)
and
their
expected
values
(
represented
as
a
line)
for
the
specified
distribution.
The
closer
the
actual
values
are
to
their
expected
values
(
i.
e.
the
more
the
actual
values
approximate
a
straight
line),
the
more
likely
the
dataset
is
of
the
specified
distribution.
The
probability
plots
for
the
UE
datasets
are
provided
in
Figures
II.
E.
2­
21
and
22.
For
the
hand
wand
inhalation
UE
dataset,
several
low
values
result
in
a
small
SW
p­
value.
The
other
dataset
is
reasonably
approximated
by
a
lognormal
distribution.
Dermal
UE
Values
(
mg/
lb
ai)
Inhalation
UE
Values
(
mg/
lb
ai)
25
0.004
52
0.005
129
0.004
27
0.004
348
0.005
56
0.005
118
0.004
176
0.016
44
0.009
41
0.016
46
0.004
15
0.004
36
0.004
83
0.004
78
0.025
78
0.012
46
0.004
36
0.022
25
0.004
63
0.018
Application
Method
Exposure
Route
Unit
Exposure
Distribution
(
mg/
lb
ai)
Shapiro­
Wilk
p­
value
Dermal
LN(
74,
64)
0.7478
Hand
Wand
Sprayer
Inhalation
LN(
0.0089,
0.0102)*
0.0005
NOTES:
LN(
µ
,
 )
represents
a
lognormal
distribution
with
mean=
µ
and
standard
deviation=
 .
*
The
mean
and
standard
deviation
represent
MLE­
based
estimates.
II.
E.
2
­
Page
30
of
46
Figure
II.
E.
2­
21
Lognormal
Probability
Plot
of
Hand
Wand
Sprayer
Dermal
UE
Data
(
MRID
#
44518501)

ln_
hw_
trees
2
3
4
5
6
7
.01
.05.10
.25
.50
.75
.90.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
22
Lognormal
Probability
Plot
of
Hand
Wand
Sprayer
Inhalation
UE
Data
(
MRID
#
44518501)

ln_
hw_
trees_
inh
1
2
3
4
.01
.05.10
.25
.50
.75
.90.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
For
the
hand
wand
inhalation
UE
dataset,
13
out
of
20
samples
were
reported
as
approximately
the
same
value.
All
13
samples
from
the
inhalation
monitors
were
reported
as
0.07
µ
g
with
slightly
different
amounts
of
active
ingredient
handled
by
the
study
subjects,
which
resulted
in
slightly
different
UE
(
mg/
lb
ai)
values.
The
value
0.07
µ
g
was
assumed
to
be
half
the
LOQ.
The
mean
and
standard
deviation
estimated
for
hand
wand
inhalation
UE
dataset
are
based
on
maximum
likelihood
estimation
(
MLE)
procedures
assuming
the
dataset
represents
a
sample
from
a
censored
lognormal
distribution.

Residue
Data
Study
Summaries
MRID
451751­
02
(
Carbaryl
Olive
DFR
Study):
The
field
phase
of
this
study
was
conducted
at
a
single
site
near
Terra
Bella,
California
which
is
in
a
major
growing
region
for
olives.
The
field
phase
of
the
study
was
conducted
during
the
period
from
November
2
to
November
17,
1998.
Sample
analyses
were
completed
by
January,
1999.
A
typical
airblast
sprayer
was
used
to
make
a
single
application
of
Sevin
XLR
Plus,
a
liquid
flowable
formulation,
at
an
application
rate
of
7.65
lb
ai/
acre.
Spray
volume
was
758
gallons
of
water
per
acre.
The
olive
trees
were
approximately
20
feet
tall
and
were
spaced
approximately
28
feet
within
each
row
while
the
rows
were
spaced
28
feet
apart
(
i.
e.,
~
56
trees/
acre).
No
significant
precipitation
was
observed
in
this
study
until
at
least
7
days
after
application.

Triplicate
DFR
samples
were
collected
out
to
14
days
after
application
using
the
Iwata
method
(
i.
e.,
a
total
surface
area
sampled
of
400
cm2/
sample
collected
with
a
1
inch
diameter
Birkestrand
leaf
punch
and
dislodged
with
a
0.01
percent
Aerosol
solution).
There
were
still
measurable
residues
14
days
after
II.
E.
2
­
Page
31
of
46
application.
The
percent
transferability
of
the
0
day
sample
was
3.6
percent
of
the
application
rate.

Statistical
Details
Dislodgeable
foliar
residue
(
DFR)
values
are
assumed
to
degrade
exponentially
over
time
(
i.
e.
degrade
by
a
constant
proportion
for
any
given
time
interval).
In
order
to
estimate
the
initial
DFR
value
(
i.
e.
DFR
value
at
day
zero)
and
the
halflife
of
the
liquid
formulation
of
carbaryl,
the
natural
logarithms
of
the
30
(
3
samples
X
10
days)
individual
DFR
samples
(
Table
II.
E.
2­
20)
from
the
olive
study
were
linearly
regressed
on
the
day
of
sample
collection.
The
form
of
the
linear
regression
is
given
below.

t
 
 
ln(
y)
1
0
+
=

The
linear
regression
parameters
were
then
used
to
calculate
initial
DFR
value
(
A0)
and
the
half­
life
(
T1/
2)
using
formulae
given
below.

0
0
 
A
=

1
 
ln(
2)
T
2
1
 
=
II.
E.
2
­
Page
32
of
46
Table
II.
E.
2­
20
Liquid
Formulation
DFR
Data
(
MRID
45175102)
Used
for
Fruit
Tree
Scenarios
Day
DFR
Values
(
mg/
cm2)

0.0035
0.0027
0
0.0030
0.0028
0.0023
1
0.0028
0.0027
0.0024
2
0.0025
0.0042
0.0029
3
0.0027
0.0028
0.0024
4
0.0028
0.0023
0.0019
5
0.0018
0.0026
0.0023
6
0.0022
0.0028
0.0023
7
0.0020
0.0012
0.0010
10
0.0010
0.0009
0.0008
14
0.0007
Transfer
Coefficient
Data
Study
Summaries
MRID
45480302
(
Hand
Pruning
Apples):
This
study
was
conducted
with
volunteer
workers
pruning
commercially
grown
apple
trees
after
two
treatments
with
the
surrogate
active
ingredient
(
ai)
carbaryl,
formulated
as
the
flowable
insecticide
Sevin
®
XLR
PLUS.
Dermal
and
inhalation
data
were
collected,
together
with
concurrent
dislodgeable
foliar
residue
(
DFR)
data.
Potential
exposures
were
measured
using
whole­
body
dosimeters
(
outer
and
inner
dosimetry),
hand
washes,
and
face/
neck
wipes
for
dermal
exposure
and
personal
II.
E.
2
­
Page
33
of
46
sampling
pumps
for
inhalation
exposure.
Transfer
coefficients
for
potential
and
total
dermal
exposure
were
calculated.

Statistical
Details
Distributional
parameters
were
estimated
for
the
adult
TC
transfer
coefficient
(
TC)
values
(
Table
II.
E.
2­
21)
from
the
apple
pruning
study.
TC
values
were
expressed
as
square
centimeters
per
hour.
Adult
TCs
were
assumed
to
be
lognormally
distributed
(
i.
e.
fitted
with
a
lognormal
distribution).
For
the
TC
dataset,
the
shape
(
 )
and
scale
(
 )
lognormal
parameters
were
estimated
by
calculating
the
mean
and
standard
deviation
of
the
natural
logarithms
(
base
e)
of
the
TCs.
Parametric
estimates
of
the
arithmetic
mean
(
µ
)
and
standard
deviation
(
 )
of
the
lognormal
distribution
were
then
calculated
based
on
the
shape
and
scale
parameter
estimates.
The
formulae
used
to
calculate
the
mean
and
standard
deviation
are
given
below.

)
 
exp(
 
µ
2
2
1
+
=

1
)
exp(
 
µ
 
2
 
=

Shapiro­
Wilk
(
S­
W)
normality
test
statistics
were
used
to
assess
the
lognormal
assumption
implicit
in
the
parametric
calculations
of
the
mean
and
standard
deviation.
The
mean,
standard
deviation,
and
p­
value
of
the
S­
W
statistic
are
provided
in
Table
II.
E.
2­
22.
A
small
p­
value
indicates
that
logarithms
of
the
TCs
are
not
normally
distributed,
or
equivalently,
that
the
TCs
are
not
lognormally
distributed.
For
the
adult
TC
dataset,
the
S­
W
p­
value
is
greater
than
0.05.

Table
II.
E.
2­
21
Liquid
Formulation
TC
Data
(
MRID
45480302)
Used
for
Fruit
Tree
Scenarios
TC
Values
(
cm2/
hr)

1119
920
903
787
534
1421
1316
940
1217
740
821
928
831
1020
606
II.
E.
2
­
Page
34
of
46
Table
II.
E.
2­
22
Lognormal
Distribution
of
TCs
Used
for
Fruit
Tree
Scenarios
Additionally,
a
probability
plot
was
used
to
qualitatively
assess
the
appropriateness
of
the
lognormal
assumption.
Generally
a
probability
plot
displays
the
actual
values
of
a
dataset
(
represented
as
points)
and
their
expected
values
(
represented
as
a
line)
for
the
specified
distribution.
The
closer
the
actual
values
are
to
their
expected
values
(
i.
e.
the
more
the
actual
values
approximate
a
straight
line),
the
more
likely
the
dataset
is
of
the
specified
distribution.
The
probability
plot
for
the
TC
dataset
is
provided
in
Figure
II.
E.
2­
23.
The
probability
plot
indicates
that
the
TC
dataset
is
reasonably
approximated
by
a
lognormal
distribution.

Figure
II.
E.
2­
23
Lognormal
Probability
Plot
of
Liquid
Formulation
TC
Data
(
MRID
45480302)

ln_
ap_
TC
6
6.5
7
7.5
.01
.05.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
5.
Ornamental
Garden­
Snail
and
Slug
Bait
Scenario
Unit
Exposure
Data
Study
Summaries
MRID
453334­
01
(
Dermal
and
Inhalation
Exposure
to
Disulfoton
Resulting
from
Residential
Application
to
Shrubs
and
Flower
Beds):
The
purpose
of
this
study
was
to
quantify
potential
dermal
(
forearm
and
hand)
and
inhalation
exposure
for
residential
applicators
a
granular
disolfoton
formulation,
which
contains
1.04
percent
disulfoton
as
the
active
ingredient.
The
maximum
Application
Method
Exposure
Route
Population
Transfer
Coefficient
Distribution
(
cm2/
hr)
Shapiro­
Wilk
p­
value
Hand
Wand
Sprayer
Dermal
Adult
LN(
943,
258)
0.9300
NOTES:
LN(
µ
,
 )
represents
a
lognormal
distribution
with
mean=
µ
and
standard
deviation=
 .
II.
E.
2
­
Page
35
of
46
application
rate
for
flower
beds
(
4
ounces
formulated
product
per
12
square
feet)
and
for
shrubs,
which
includes
rosebushes,
(
4
ounces
formulated
product
per
1
foot
shrub
height)
was
used
in
this
study.

The
field
study
was
conducted
at
the
Bayer
Corporation
Research
Farm,
Vero
Beach,
Florida.
A
total
of
15
volunteers
were
monitored
using
passive
dosimetry
(
hand/
forearm
wash
solutions
and
personal
air
monitors).
Application
of
the
product
was
made
by
pouring
the
granules
into
the
measuring
cup/
lid
attached
to
the
product
package,
and
then
distributing
the
granules
onto
the
soil
around
the
base
of
a
shrub
or
onto
a
flower
bed.
The
granules
were
then
soil­
incorporated
with
a
garden
rake.
Each
volunteer
applied
granular
disulfoton
around
shrubs
while
wearing
gloves
and
then
again
without
gloves.
A
total
of
60
(
i.
e.,
15
volunteers
x
4
exposure
scenarios)
replicates
were
monitored.
Only
exposure
data
from
the
30
replicates
who
did
not
wear
gloves
were
reported.
The
test
site
was
a
fallow
test
field,
approximately
1
acre
in
size.
Two
sets
of
sub­
plots
were
established:
(
1)
shrub
test­
plots,
each
containing
10
oleander
shrubs
(
approximately
48
inches
high);
and
(
2)
flower­
bed
sub­
plots,
each
containing
simulated
plants,
(
e.
g.,
12
to
14
inch
high
stakes
placed
on
approximately
24
inch
centers).

Each
volunteer
applied
approximately
10
pounds
of
formulated
product
per
application.
Shrubs
were
treated
by
spreading
16
ounces
of
granules
(
i.
e.,
4
ounces
per
1
foot
of
shrub)
in
a
circle
around
each
shrub's
base.
The
granules
were
then
incorporated
into
the
top
1­
2
inches
of
soil
using
a
new
garden
rake.
Flower
beds
were
treated
by
sprinkling
4
ounces
of
granules
to
each
12
square
feet
of
a
total
480
square
feet
area,
and
incorporating
the
product
into
the
top
1­
2
inches
of
soil
using
a
new
garden
rake.

All
of
the
inhalation
exposure
data
were
either
non­
detect
or
less
than
the
limit
of
quantitation
(
LOQ).
Most
of
the
hand/
forearm
dermal
washing
samples
returned
results
greater
than
the
LOQ.
The
author
reported
that
the
time
it
took
to
treat
shrubs
ranged
between
18
and
29
minutes.
The
time
that
it
took
to
treat
flowerbeds
ranged
between
20
and
40
minutes.

Statistical
Details
Distributional
parameters
were
estimated
for
the
dermal
unit
exposure
(
UE)
values
(
Table
II.
E.
2­
23)
for
the
granular
formulation
of
chemical
H
based
on
surrogate
chemical
data.
Dermal
UE
values
represent
milligrams
exposure
per
pound
of
active
ingredient
of
a
pesticide
handled.
All
UEs
were
assumed
to
be
lognormally
distributed
(
i.
e.
fitted
with
a
lognormal
distribution).
For
the
dataset,
the
shape
(
 )
and
scale
(
 )
lognormal
parameters
were
estimated
by
calculating
the
mean
and
standard
deviation
of
the
natural
logarithms
(
base
e)
of
the
UEs.
Parametric
estimates
of
the
arithmetic
mean
(
µ
)
and
standard
deviation
(
 )
of
the
lognormal
distribution
were
then
calculated
based
on
the
shape
and
scale
II.
E.
2
­
Page
36
of
46
parameter
estimates.
The
formulae
used
to
calculate
the
mean
and
standard
deviation
are
given
below.

)
 
exp(
 
µ
2
2
1
+
=

1
)
exp(
 
µ
 
2
 
=

The
Shapiro­
Wilk
(
S­
W)
normality
test
statistic
was
used
to
assess
the
lognormal
assumption
implicit
in
the
parametric
calculations
of
the
mean
and
standard
deviation.
The
mean,
standard
deviation,
and
p­
value
of
the
S­
W
statistic
are
provided
in
Table
II.
E.
2­
24.
A
small
p­
value
indicates
that
logarithms
of
the
UEs
are
not
normally
distributed,
or
equivalently,
that
the
UEs
are
not
lognormally
distributed.
The
granular
dermal
UE
dataset
resulted
in
an
S­
W
statistic
with
a
pvalue
less
than
0.05.

Table
II.
E.
2­
23
Granular
UE
Data
(
MRID
45333401)
Used
for
Ornamental
Garden
Scenarios
Table
II.
E.
2­
24
Lognormal
Distribution
of
UEs
Used
for
Ornamental
Garden
Scenarios
Dermal
UE
Values
(
mg/
lb
ai)
0.001
0.029
0.019
0.206
0.085
0.001
0.001
0.001
0.042
0.050
0.012
0.032
0.119
0.001
0.082
Application
Method
Exposure
Route
Unit
Exposure
Distribution
(
mg/
lb
ai)
Shapiro­
Wilk
p­
value
Granular
Dermal
LN(
0.23,
5.8)*
0.0087
NOTES:
LN(
µ
,
 )
represents
a
lognormal
distribution
with
mean=
µ
and
standard
deviation=
 .
*
The
mean
and
standard
deviation
represent
MLE­
based
estimates.
II.
E.
2
­
Page
37
of
46
Additionally,
a
probability
plot
was
used
to
qualitatively
assess
the
appropriateness
of
the
lognormal
assumption.
Generally
a
probability
plot
displays
the
actual
values
of
a
dataset
(
represented
as
points)
and
their
expected
values
(
represented
as
a
line)
for
the
specified
distribution.
The
closer
the
actual
values
are
to
their
expected
values
(
i.
e.
the
more
the
actual
values
approximate
a
straight
line),
the
more
likely
the
dataset
is
of
the
specified
distribution.
The
probability
plot
for
the
UE
dataset
is
provided
in
Figure
II.
E.
2­
24.
For
the
granular
dermal
UE
dataset,
several
low
values
result
in
a
small
S­
W
p­
value.

Figure
II.
E.
2­
24
Lognormal
Probability
Plot
of
Granular
Dermal
UE
Data
(
MRID
45333401)

ln_
dis_
gran_
fb
­
1
0
1
2
3
4
5
6
.01
.05
.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
For
the
granular
dermal
UE
dataset,
5
out
of
15
samples
were
reported
as
half
the
LOQ.
The
mean
and
standard
deviation
estimated
for
hand
wand
dermal
UE
dataset
are
based
on
maximum
likelihood
estimation
(
MLE)
procedures
assuming
the
dataset
represents
a
sample
from
a
censored
lognormal
distribution.
Since
all
inhalation
samples
were
either
non­
detect
or
less
than
the
LOQ,
half
the
LOQ
(
0.00001
mg/
lb
aihandled)
was
used
as
an
estimate
of
inhalation
UE.

6.
Indoor
Crack
and
Crevice
Scenario
Unit
Exposure
Data
Study
Summaries
MRID
410547­
05
(
Exposures
of
Applicators
to
Propoxur
during
Residential
Application
of
an
Aerosol
Spray
Containing
1%
Propoxur):
Applicators
in
the
study
each
applied
one
16­
ounce
aerosol
can
in
each
of
the
15
residences
situated
in
Vero
Beach,
Florida.
The
entire
contents
were
applied
to
each
house.
The
volunteers
sprayed
to
cracks,
crevices
along
baseboards
and
other
II.
E.
2
­
Page
38
of
46
woodwork,
under
sinks
and
behind
appliances.
The
majority
of
the
exposure
was
to
the
hands,
neck
and
head
(~
85).

Statistical
Details
Distributional
parameters
were
estimated
for
the
(
dermal)
transferable
residue
(
Table
II.
E.
2­
25)
and
(
inhalation)
air
concentration
(
Table
II.
E.
2­
26)
values
for
pressurized
can
sprayer
applications
of
Chemical
G.
Transferable
residue
(
TR)
values
represent
milligrams
exposure
per
square
centimeter
and
air
concentration
(
AC)
values
represent
milligrams
per
cubic
meter.
The
average
TR
values
across
houses
were
assumed
to
be
lognormally
distributed
(
i.
e.
fitted
with
a
lognormal
distribution);
whereas
the
individual
AC
values
within
a
house
were
assumed
to
be
lognormally
distributed.
For
TR
and
AC
datasets,
the
shape
(
 )
and
scale
(
 )
lognormal
parameters
were
estimated
by
calculating
the
mean
and
standard
deviation
of
the
natural
logarithms
(
base
e)
the
TR
and
AC
values,
respectively.
Parametric
estimates
of
the
arithmetic
mean
(
µ
)
and
standard
deviation
(
 )
of
the
lognormal
distribution
were
then
calculated
based
on
the
shape
and
scale
parameter
estimates.
The
formulae
used
to
calculate
the
mean
and
standard
deviation
are
given
below.

)
 
exp(
 
µ
2
2
1
+
=

1
)
exp(
 
µ
 
2
 
=

Shapiro­
Wilk
(
S­
W)
normality
test
statistics
were
used
to
assess
the
lognormal
assumption
implicit
in
the
parametric
calculations
of
the
mean
and
standard
deviation.
The
means,
standard
deviations,
and
p­
values
of
the
S­
W
statistics
are
provided
in
Table
II.
E.
2­
27.
A
small
p­
value
indicates
that
logarithms
of
the
TRs
(
or
the
ACs)
are
not
normally
distributed,
or
equivalently,
that
the
TRs
(
or
ACs)
are
not
lognormally
distributed.
For
both
TR
(
dermal)
and
AC
(
inhalation)
pressurized
can
datasets,
the
S­
W
p­
values
are
greater
than
0.05.
II.
E.
2
­
Page
39
of
46
Table
II.
E.
2­
25
Deposition
Data
(
MRID
41054705)
Used
for
Crack
and
Crevice
Scenarios
Hard
Surfaces
Deposition
Values
(
mg/
cm2)
0.590
0.414
0.020
0.002
0.080
Table
II.
E.
2­
26
Air
Concentration
Data
(
MRID
41054705)
Used
for
Crack
and
Crevice
Scenarios
Air
Concentration
Values
(
mg/
m3)
5.3
5.8
5.3
5.4
4.2
3.4
3.9
2.4
4.0
2.4
6.5
5.7
7.5
6.9
5.4
6.2
8.6
8.6
2.5
4.4
Table
II.
E.
2­
27
Lognormal
Distributions
of
Surface
Deposition
and
Air
Concentrations
Used
for
Crack
and
Crevice
Scenarios
Additionally,
probability
plots
were
used
to
qualitatively
assess
the
appropriateness
of
the
lognormal
assumptions.
Generally
a
probability
plot
displays
the
actual
values
of
a
dataset
(
represented
as
points)
and
their
expected
values
(
represented
as
a
line)
for
the
specified
distribution.
The
closer
the
actual
values
are
to
their
expected
values
(
i.
e.
the
more
the
actual
values
approximate
a
straight
line),
the
more
likely
the
dataset
is
of
the
specified
distribution.
The
probability
plots
for
the
TR
and
AC
datasets
are
provided
in
Figures
II.
E.
2­
25
and
26.
The
probability
plots
indicate
that
both
datasets
are
reasonably
approximated
by
lognormal
distributions.
Application
Method
Exposure
Route
Deposition
(
mg/
cm2)
and
Air
Concentration
(
mg/
m3)
Distributions
Shapiro­
Wilk
p­
value
Dermal
LN(
0.0010,
0.0185)
0.5962
Handheld
Compression
Sprayer
Inhalation
LN(
0.0053,
0.0021)
0.1770
NOTES:
LN(
µ
,
 )
represents
a
lognormal
distribution
with
mean=
µ
and
standard
deviation=
 .
II.
E.
2
­
Page
40
of
46
Figure
II.
E.
2­
25
Lognormal
Probability
Plot
of
Deposition
Data
(
MRID
41054705)

ln_
propx_
hard_
avg
­
7
­
6
­
5
­
4
­
3
­
2
­
1
0
1
.01
.05
.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Figure
II.
E.
2­
26
Lognormal
Probability
Plot
of
Air
Concentration
Data
(
MRID
41054705)

ln_
propx_
hard_
inh
0.8
1
1.2
1.4
1.6
1.8
2
2.2
.01
.05
.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Residue
Data
Study
Summaries
MRID
410547­
03
(
Exposure
to
Propoxur
for
Residents
of
Homes
Treated
with
BAYGON
70%
WP):
A
formulation
of
Propoxur,
BAYGON
70WP,
was
applied
as
a
coarse
spray
to
cracks
and
crevices,
baseboards,
and
other
small
areas
commonly
treated
for
insect
control
using
a
handheld
compression
sprayer.
An
average
of
1.2
ounces
(
0.7
­
1.8
)
of
ai
was
applied
to
each
of
the
five
houses
monitored.
Applications
took
2­
to
34
minutes
to
complete.

Surface
residues
and
air
levels
of
propoxur
were
measured
at
intervals
of
up
to
48
hours
after
treatment.
Five
types
of
surfaces
were
evaluated.
The
media
were
distributed
in
the
rooms
prior
to
treatment.
Triplicate
samples
of
each
medium
were
collected,
before
treatment,
immediately
after
applications,
and
at
intervals
of
6,
12,
24
and
48
hours
post­
application.
In
the
kitchens,
vinyl
tile
squares
were
placed
on
the
floor
and
on
counter
tops.
The
living
rooms
and
bedrooms
were
sampled
using
squares
of
nylon
carpet
with
a
1
cm
nap
placed
on
the
floor
and
fabric
squares
were
located
on
the
furniture.
These
samples
were
used
for
total
residue
analysis.
Air
concentrations
of
propoxur
were
determined
by
drawing
air,
at
a
rate
of
1
L/
minute,
through
a
sampling
apparatus
whose
inlet
was
located
12
inches
above
the
floor.
All
sampling
periods
were
at
least
one
hour.
II.
E.
2
­
Page
41
of
46
7.
Pet
Collar
Exposure
Residue
Data
Study
Summaries
MRID
#
45792201
(
Carbaryl
Pet
Flea
Collar
Study):
Sixteen
dogs
were
fitted
with
16%
carbaryl
flea
collars.
The
dogs
were
of
various
breeds
and
sizes.
All
dogs
that
had
previously
worn
any
type
of
flea
collar
or
had
been
in
contact
with
any
flea
killing
material
were
thoroughly
washed
prior
to
placement
of
the
collars.
All
dogs
were
allowed
to
follow
their
normal
daily
habits.
The
dogs
were
petted
at
24
hours,
96
hours
and
1
week.
During
each
petting
session,
each
dog
was
petted
for
one
10
minute
period
followed
by
another
10
minute
period,
each
time
petters
wore
a
new
pair
of
pre­
washed
white
cotton
gloves.
All
dogs
were
petted
in
such
a
manner
that
contact
with
all
portions
of
their
bodies
resulted.
The
four
gloves
from
each
20
minute
session
were
considered
a
single
sample
and
were
placed
in
individual
plastic
bags
for
analysis.

Transfer
Coefficient
Data
Study
Summaries
MRID
446584­
01
(
Carbaryl
Dog
Groomer
Study):
The
data
collected
reflect
the
dermal
and
respiratory
exposure
of
commercial
pet
groomers
applying
the
end
use
product,
Adams
®
Carbaryl
Flea
and
Tick
Shampoo
containing
0.50
percent
carbaryl.
In
this
study,
applications
of
Adams
®
Carbaryl
Flea
and
Tick
Shampoo
were
made
by
professional
pet
groomers
to
8
dogs
at
2
sites
in
Georgia.
A
total
of
16
replicates
were
monitored
for
dermal
and
inhalation
exposure.
Eight
dogs
of
various
sizes
and
hair
lengths
were
shampooed
during
each
replicate.
Dermal
exposure
was
monitored
with
face
and
neck
swabs,
100
percent
cotton
union
suit
dosimeter
worn
underneath
a
short­
sleeved
t­
shirt,
long
pants
and
a
65/
35
polyester
cotton
long­
sleeved
smock
(
i.
e.,
represents
a
short­
sleeved
shirt
under
a
long­
sleeved
coat/
smock).
Hand
exposure
was
quantified
using
handwash
rinses
(
no
protective
gloves
were
worn).
Inhalation
exposure
was
monitored
using
personal
air
pumps
with
XAD2
resin
tubes.

The
dogs
were
wetted,
shampooed
to
a
lather
(
lather
remained
on
dogs
for
5
minutes)
and
rinsed.
After
completing
8
dog
shampoos
the
dosimeters
were
collected.
Face/
neck
swabs
and
2
hand
rinses
were
performed
along
with
collection
of
the
100
percent
cotton
union
suit.
II.
E.
2
­
Page
42
of
46
Table
II.
E.
2­
28
TC
Data
Used
for
Pet
Collar
Scenario
(
Empirical
Distribution)
1
Groomer
µ
g
exposure
Duration:
hours
µ
g/
hour
ai
deposited
µ
g/
cm2*
Dislodged:
2.97
%
efficiency
assumed
µ
g/
cm2
Transfer
Coefficient
(
adults)
cm2/
hour
Transfer
Coefficient
(
children)
cm2/
hour
/
3
8796
2.88
3054
37.5
1.114
2742
1016
6199
2.58
2403
31.0
0.921
2610
967
1408
3.07
459
18.6
0.552
831
308
2914
2.48
1175
36.4
1.081
1087
403
5667
3.08
1840
32
0.950
1936
717
2527
3.18
795
19
0.564
1409
522
2,348
2.93
801
15.9
0.472
1696
628
2961
2.72
1089
7.75
0.230
4731
1752
1135
4.03
282
14.8
0.440
642
238
14872
3.88
3833
28.8
0.855
4481
1660
1026
3.17
324
16.6
0.493
657
243
13490
4.05
3331
56.98
1.692
1968
729
4275
4.92
869
25
0.743
1170
433
4461
3.45
1293
42.25
1.255
1030
382
1511
3.03
499
8.87
0.263
1894
702
777
3.00
259
48.6
1.443
179
66
Average
1817
673
1
Source
Carbaryl
Groomer
Exposure
Study
(
activity
­
wash/
dip/
groom).
Each
vet
tech
treated/
handled
8
dogs:
held
small
dogs
w/
arms
and
torso;
some
dogs
climbed
on
person's
shoulders
while
grooming
etc.
2
Average
transfer
efficiency
2.97%
=(
powder
(
0.62%)
+
aerosol
(
3.3%)
+
pump
spray
(
5%))/
3;
.
3
The
transfer
coefficients
derived
from
this
study
were
adjusted
by
an
allometric
scaling
factor
based
on
the
relative
size
of
children
to
adults
to
derive
an
appropriate
transfer
coefficient
for
children
Adult:
Child
surface
area
ratio
­
2.7:
1
(
avg.
Adult
3169:
avg
child
1174)
*
The
amount
ai
per
dog
was
measured
in
the
study
along
with
the
animal's
weight.
The
suface
areas
of
the
dogs
were
estimated
using
an
equation
for
estimating
mammal
surface
area
described
in
the
Wildlife
Exposure
Factors
Handbook.

8.
Golf
Course
Exposure
Residue
Data
Study
Summaries
MRID
451143­
01
(
Carbaryl
Turf
Transferable
Residue
Study):
A
TTR
study
was
conducted
at
individual
sites
in
three
states
using
the
ORETF
roller
sampling
method.
The
data
used
in
this
assessment
was
from
the
Georgia
site.
Bermudagrass
was
the
variety
of
turfgrass
treated
at
the
Georgia
site.
Field
work
took
place
over
three
week
intervals
at
each
site.
Applications
were
made
and
samples
were
collected
essentially
in
October
of
1998
Georgia.
Two
applications
were
made
7
days
apart
at
each
site.
All
applications
in
this
study
were
completed
at
a
rate
of
8.17
lb
ai/
acre.
Applications
were
made
with
typical
groundboom
sprayers
using
approximately
55
and
31
gallons
of
water
per
acre,
respectively.
All
applications
were
made
using
Dragon
Sevin
Liquid
which
is
a
II.
E.
2
­
Page
43
of
46
flowable
concentrate
formulation
that
contains
carbaryl
at
a
nominal
concentration
of
21
percent
by
weight
or
2
lb
ai/
gallon.

There
was
approximately
from
1
inch
up
to
2.7
inches
of
irrigation
water
on
the
day
of
the
final
application.
Additionally,
on
the
day
of
the
final
application,
rain
was
noted
that
ranged
in
accumulations
of
0.36
inches.
Mowing
events
were
not
noted
in
the
data
from
the
Georgia
site.
Triplicate
TTR
samples
were
collected
using
the
ORETF
roller
method
at
8
intervals
out
to
14
days
after
the
last
application.
All
but
two
samples
were
collected
during
the
1st
week
of
the
study.
In
all
cases,
residue
levels
exceeded
the
LOQ
at
14
days
after
application.

Statistical
Details
Turf
transferable
residues
(
TTR)
values
are
assumed
to
degrade
exponentially
over
time
(
i.
e.
degrade
by
a
constant
proportion
for
any
given
time
interval).
In
order
to
estimate
the
initial
TTR
value
(
i.
e.
TTR
value
at
day
zero)
and
the
halflife
of
the
liquid
formulation
of
Carbaryl,
the
natural
logarithms
of
the
27
(
3
samples
X
9
days)
individual
TTR
samples
(
Table
II.
E.
2­
29)
from
the
Georgia
site
were
linearly
regressed
on
the
day
of
sample
collection.
The
form
of
the
linear
regression
is
given
below.

t
 
 
ln(
y)
1
0
+
=

The
linear
regression
parameters
were
then
used
to
calculate
initial
TTR
value
(
A0)
and
the
half­
life
(
T1/
2)
using
formulae
given
below.

0
0
 
A
=

1
 
ln(
2)
T
2
1
 
=
II.
E.
2
­
Page
44
of
46
Table
II.
E.
2­
29
Liquid
Formulation
TTR
Data
(
MRID
#
45114301)
Used
for
Carbaryl
Golf
Scenarios
Transfer
Coefficient
Data
Study
Summaries
Ballee,
1990
(
chlorothalonil
TC
data)
and
Moran
et
al.,
1987
(
flurprimidol
TC
data):
The
data
used
to
derive
transfer
coefficients
were
based
on
two
measurements
of
four
individuals
playing
golf
on
two
golf
courses
treated
with
chlorothalonil
(
Ballee,
1990),
and
the
exposure
of
golfers
(
four
volunteers)
to
flurprimidol
(
Moran
et
al.,
1987).
For
both
studies,
an
assumed
transfer
efficiency
of
1%
was
used
to
calculate
the
transfer
coefficients,
since
the
studies
were
conducted
using
sprayable
formulations.
Day
TTR
Values
(
mg/
cm2)

0.00130
0.00122
0
0.00152
0.00067
0.00073
0.5
0.00147
0.00041
0.00042
1
0.00047
0.00020
0.00028
2
0.00023
0.00014
0.00027
3
0.00050
0.00040
0.00023
5
0.00010
0.00011
0.00013
7
0.00031
0.00003
0.00015
10
0.00022
0.00007
0.00002
14
0.00015
II.
E.
2
­
Page
45
of
46
Statistical
Details
Distributional
parameters
were
estimated
for
the
adult
TC
transfer
coefficient
(
TC)
values
(
Table
II.
E.
2­
30)
from
the
Ballee
and
Moran
studies.
TC
values
were
expressed
as
square
centimeters
per
hour.
Adult
TCs
were
assumed
to
be
lognormally
distributed
(
i.
e.
fitted
with
a
lognormal
distribution).
For
the
TC
dataset,
the
shape
(
 )
and
scale
(
 )
lognormal
parameters
were
estimated
by
calculating
the
mean
and
standard
deviation
of
the
natural
logarithms
(
base
e)
of
the
TCs.
Parametric
estimates
of
the
arithmetic
mean
(
µ
)
and
standard
deviation
(
 )
of
the
lognormal
distribution
were
then
calculated
based
on
the
shape
and
scale
parameter
estimates.
The
formulae
used
to
calculate
the
mean
and
standard
deviation
are
given
below.

)
 
exp(
 
µ
2
2
1
+
=

1
)
exp(
 
µ
 
2
 
=

The
Shapiro­
Wilk
(
S­
W)
normality
test
statistic
was
used
to
assess
the
lognormal
assumption
implicit
in
the
parametric
calculations
of
the
mean
and
standard
deviation.
The
mean,
standard
deviation,
and
p­
value
of
the
S­
W
statistic
are
provided
in
Table
II.
E.
2­
31.
A
small
p­
value
indicates
that
logarithms
of
the
TCs
are
not
normally
distributed,
or
equivalently,
that
the
TCs
are
not
lognormally
distributed.
For
the
adult
TC
dataset,
the
S­
W
p­
value
is
greater
than
0.05.

Table
II.
E.
2­
30
Liquid
Formulation
TC
Data
Used
for
Golf
Scenarios
TC
Values
(
cm2/
hr)

391
329
561
547
592
533
385
508
756
522
264
278
II.
E.
2
­
Page
46
of
46
Table
II.
E.
2­
31
Lognormal
Distribution
of
TCs
Used
for
Golf
Scenarios
Additionally,
a
probability
plot
was
used
to
qualitatively
assess
the
appropriateness
of
the
lognormal
assumption.
Generally
a
probability
plot
displays
the
actual
values
of
a
dataset
(
represented
as
points)
and
their
expected
values
(
represented
as
a
line)
for
the
specified
distribution.
The
closer
the
actual
values
are
to
their
expected
values
(
i.
e.
the
more
the
actual
values
approximate
a
straight
line),
the
more
likely
the
dataset
is
of
the
specified
distribution.
The
probability
plot
for
the
TC
dataset
is
provided
in
Figure
II.
E.
2­
27.
The
probability
plot
indicates
that
the
TC
dataset
is
reasonably
approximated
by
a
lognormal
distribution.

Figure
II.
E.
2­
27
Lognormal
Probability
Plot
of
Liquid
Formulation
TC
Data
Used
for
Golf
Scenarios
ln_
golf_
TC
5
5.5
6
6.5
7
.01
.05
.10
.25
.50
.75
.90
.95
.99
­
3
­
2
­
1
0
1
2
3
Normal
Quantile
Application
Method
Exposure
Route
Population
Transfer
Coefficient
Distribution
(
cm2/
hr)
Shapiro­
Wilk
p­
value
Liquid
Broadcast
Dermal
Adult
LN(
475,
158)
0.4068
NOTES:
LN(
µ
,
 )
represents
a
lognormal
distribution
with
mean=
µ
and
standard
deviation=
 .
