1
For
example,
the
information
entered
in
the
ATP
worksheets
for
sample
#
B982305A
(
Millbrae,
CA)
lists
the
date
for
the
original
sample
density
determination
as
12­
23­
98.
This
is
also
the
date
entered
for
the
inoculation
of
the
test
media
and
for
the
comparability
study
results.
As
discussed
in
this
attachment,
however,
the
Colitag
 
method
takes
24
hours
to
complete
and
the
reference
methods
each
require
a
minimum
of
48
hours.
Hence,
the
dates
entered
in
the
worksheets
cannot
possibly
reflect
all
of
the
actual
dates
for
the
corresponding
steps
in
the
evaluation
of
the
Colitag
 
method.
EPA's
presumption
for
this
method
evaluation,
as
for
previous
method
evaluations,
is
that
the
certified
drinking
water
laboratory
performing
the
comparability
studies
employed
a
reasonable
hold
time
for
all
samples.
Although
the
analysis
described
in
this
Attachment
can
confirm
this
for
some
samples
only,
the
information
available
to
EPA
is
insufficient
to
refute
EPA's
general
presumption
for
any
of
the
samples.
Attachment
1
Description
of
CPI
Comparability
Study
Log
Sheet
Hold
Times
for
Colitag
 
Background/
Assumptions
EPA's
analysis
of
the
Colitag
 
ATP
Worksheets
suggests
that
some
data
were
likely
entered
onto
the
worksheet
(
and
were
dated)
days
after
the
original
data
collection.
1
One
explanation
for
such
cases
is
that
data
may
have
been
entered
into
laboratory
notebooks
on
a
real­
time
basis
and
subsequently
transferred
to
the
worksheets,
with
the
dates
in
the
submitted
worksheets
reflecting
the
dates
of
transcription
rather
than
those
corresponding
to
the
underlying
lab
procedures.

Laboratory
record
keeping
practices
such
as
this
(
i.
e.,
transferring
data
from
laboratory
notebooks
to
summary
worksheets)
are
not
uncommon.

Consistent
with
past
practice
under
the
ATP
program,
original
lab
notebooks
were
not
requested
by
EPA,
nor
were
they
submitted
to
EPA.
Since
such
notebooks
were
not
submitted
to
EPA,

they
are
not
part
of
the
administrative
record
for
the
decision
to
approve
Colitag
 
.
Accordingly,

any
information
contained
in
such
notebooks
did
not
enter
into
EPA's
decision.
As
a
result,
EPA
does
not
have
the
necessary
information
available
to
calculate
a
precise
"
hold
time"
(
i.
e.,
the
time
elapsed
between
the
original
[
pre­
chlorination]
density
determination
and
the
post­
chlorination
measurement,
both
of
which
are
used
to
calculate
log
reduction).
Nonetheless,
the
information
in
the
record,
coupled
with
an
understanding
of
the
timing
and
scope
of
the
test
procedures
used
in
the
comparability
studies,
does
allow
EPA
to
conservatively
estimate
hold
times.
2
Colitag
 
test
results
[
total
coliforms
+
E.
Coli]
(
24
hr)
Inoculation
of
test
media
(
presumed
to
take
place
immediately
after
the
postchlorination
measurement)
Postchlorination
measurement
BGLB
(
reference
method)
total
coliform
results
(
48­
96
hr)

EC
MUG
(
reference
method)
E.
Coli
results
(
48­
72
hr)
The
key
information
that
EPA
determined
it
could
consistently
and
reliably
use
to
evaluate
hold
time
was
the
"
Test
Material/
Equipment
Transfer
Receipt"
(
the
"
chain
of
custody"
record
documenting
the
date/
time
when
the
sample
was
received
in
the
Biovir
laboratory)
and
the
data
sheets
documenting
"
Results
of
ATP
Comparability
Study
of
Method
for
Total
Coliform
Bacteria/

Escherichia
Coli"
(
i.
e.,
the
"
comparability
study
sheets").

One
can
be
confident
that
the
initial
density
determination
occurred
no
earlier
than
the
date
that
the
sample
was
received
in
the
laboratory.
Moreover,
when
the
sample
arrived
in
the
laboratory
in
the
afternoon,
one
can
reasonably
predict
that
in
most
instances
the
initial
density
determination
occurred
no
earlier
than
the
following
day,
based
on
the
complexity
and
duration
of
the
test
procedure.
Accordingly,
EPA
believes
it
is
reasonable
to
add
1
day
to
the
laboratory
receipt
date
to
establish
a
reasonably
conservative
estimate
of
the
pre­
chlorination
measurement
date
when
the
sample
arrived
in
the
afternoon.

One
can
also
estimate
the
post­
chlorination
measurement
date,
based
on
an
understanding
of
the
comparability
study
timing,
as
outlined
below:

Comparability
Study
Timeline
Complete
total
coliform
results
should
be
available
no
later
than
4
days
after
inoculation,
since
the
2
If
EPA
were
to
build
in
an
assumption
that
any
length
of
time
elapsed
in
between
the
post­
chlorination
measurements
and
inoculation
of
the
test
media,
this
would
necessarily
push
back
the
estimated
points
at
which
the
post­
chlorination
measurements
were
taken
and
reduce
the
"
hold
time"
between
the
initial
density
determination
and
the
post­
chlorination
measurements.

3
total
coliform
reference
method
requires
48­
96
hours
to
complete.
Similarly,
complete
E.
Coli
results
should
be
available
no
later
than
3
days
after
inoculation,
since
the
E.
Coli
reference
method
requires
48­
72
hours
to
complete.
The
Colitag
 
portion
of
the
test
results
should
be
available
one
day
after
inoculation,
since
the
Colitag
 
test
requires
24
hours
to
complete
for
both
total
coliforms
and
E.
coli.
(
For
the
purposes
of
estimating
hold
time
for
these
comparability
study
samples,
EPA
has
made
the
conservative
presumption
that
inoculation
of
the
test
media
takes
place
immediately
following
the
post­
chlorination
measurement.
2
)

EPA
surmises
that
the
dates
on
the
comparability
study
sheets
reflect
either
the
date
that
data
were
initially
transferred
(
no
earlier
than
24
hours
after
inoculation
of
the
test
media,
when
Colitag
 
results
might
be
available)
or
the
date
when
all
data
were
transferred
(
no
earlier
than
48­
72
or
48­
96
hours
after
inoculation
of
the
test
media,
when
results
from
the
E.
coli
and
total
coliform
reference
methods,
respectively,
might
be
available).
Accordingly,
one
may
subtract
1­
3
days
from
the
date
on
the
comparability
study
sheets
for
E.
Coli
results
and
1­
4
days
from
the
date
on
the
study
sheets
for
total
coliform
results,
as
appropriate,
to
identify
a
range
of
conservatively­
estimated
post­
chlorination
measurement
dates.
As
reflected
in
the
estimates
of
sample­
specific
hold
times
(
below),
in
those
cases
where
the
higher
end
of
the
range
yielded
an
illogical
estimated
date
(
i.
e.,
suggestive
of
the
post­
chlorination
measurement
preceding
the
prechlorination
measurement)
EPA
has
used
a
narrower
range
for
the
study­
duration
adjustment
(
reflecting
a
more
conservative
approach)
to
yield
a
reasonable
estimated
date.

A
conservative
hold­
time
may
be
computed
as
the
difference
between
the
estimated
prechlorination
and
post­
chlorination
measurement
dates.
Applying
this
approach,
EPA
concluded
that
reasonable
hold­
times
could
be
established
for
the
six
samples
described
below.
Applying
the
approach
to
the
other
four
samples
(
Shaumberg,
IL
990052A,
Mission,
KS
990217A,

Jacksonville,
FL
990025A
and
Salem,
OR
990273A)
did
not
allow
EPA
to
establish
a
reasonable,
4
conservative
hold­
time
estimate.
EPA
hastens
to
add,
however,
that
this
information
does
not
lead
inexorably
to
the
conclusion
that
the
hold
times
were
unreasonable
and
EPA
continues
to
make
the
general
assumption
that
each
of
the
samples
was
held
for
a
reasonable
period
of
time.

In
the
alternative,
EPA
investigated
what
impact
removal
of
the
four
samples
mentioned
above
would
have
on
the
Agency's
assessment
of
Colitag
 
.
As
is
described
in
Attachment
2,

elimination
of
these
four
samples
from
the
statistical
analysis
would
not
alter
EPA's
conclusion.

Sample­
Specific
Hold­
time
Estimates
Millbrae,
CA
982084A.
This
sample
was
received
at
10:
35
on
11­
10­
98
as
is
indicated
on
the
Biovir
Laboratories
Inc.
Sample
Transfer
Receipt.
Because
this
sample
was
received
early
in
the
day,
EPA
assumed
that
work
began
on
this
sample
on
11­
10­
98.
Thus,
11­
10­
98
was
established
as
the
estimated
pre­
chlorination
measurement
date.
The
date
on
the
study
sheets
for
this
sample
was
11­
11­
98
(
for
both
the
total
coliform
and
E.
coli
results).
In
light
of
the
above
discussion,

and
considering
that
post­
chlorination
measurements
cannot
precede
pre­
chlorination
measurements,
an
adjustment
of
one
day
is
reasonable,
in
which
case
the
estimated
postchlorination
measurement
date
is
11­
10­
98.
Therefore,
the
estimated
hold
time
for
this
sample
is
<
24
hours.

Millbrae,
CA
982305A.
This
sample
was
received
at
14:
45
on
12­
16­
98
as
is
indicated
on
the
Biovir
Laboratories
Inc.
Sample
Transfer
Receipt.
Because
this
sample
arrived
late
in
the
day,

EPA
assumed
that
work
began
on
this
sample
on
12­
17­
98.
Thus,
12­
17­
98
was
established
as
the
estimated
pre­
chlorination
measurement
date.
The
date
on
the
study
sheets
for
this
sample
was
12­
23­
98
(
for
both
the
total
coliform
and
E.
coli
results).
In
light
of
the
above
discussion,
an
adjustment
of
1­
3
days
is
conservative,
in
which
case
the
estimated
post­
chlorination
measurement
date
is
between
12­
20­
98
and
12­
22­
98.
Therefore,
the
estimated
hold
time
for
this
sample
is
no
more
than
3­
5
days.

Watertown,
WI
990095A.
This
sample
was
received
at
09:
00
on
2­
2­
99
as
is
indicated
on
the
5
Biovir
Laboratories
Inc.
Sample
Transfer
Receipt.
Because
this
sample
arrived
early
in
the
day,

EPA
assumed
that
work
began
on
this
sample
on
2­
2­
99.
Thus,
2­
2­
99
was
established
as
the
estimated
pre­
chlorination
measurement
date.
The
dates
on
the
total
coliform
study
sheet
for
this
sample
was
2­
7­
99.
In
light
of
the
above
discussion
an
adjustment
of
1­
4
days
is
conservative,
in
which
case
the
estimated
post­
chlorination
measurement
date
is
between
2­
3­
99
and
2­
6­
99.

Therefore,
the
estimated
hold
time
for
this
sample
is
no
more
than
1­
4
days.
(
Using
the
E.
coli
study
sheet
date
of
2­
6­
99
and
the
conservative
adjustment
of
1­
3
days
for
E.
coli
results
yields
an
estimated
hold
time
of
no
more
than
1­
3
days.)

Mission,
KS,
990442A
This
sample
was
received
at
13:
15
on
4­
27­
99
as
is
indicated
on
the
Biovir
Laboratories
Inc.
Sample
Transfer
Receipt.
Because
this
sample
arrived
late
in
the
day,

EPA
assumed
that
work
began
on
this
sample
on
4­
28­
99.
Thus,
4­
28­
99
was
established
as
the
estimated
pre­
chlorination
measurement
date.
The
date
on
the
total
coliform
study
sheet
for
this
sample
was
4­
30­
99
(
no
date
is
available
for
the
E.
coli
results).
In
light
of
the
above
discussion,

and
considering
that
post­
chlorination
measurements
cannot
precede
pre­
chlorination
measurements,
an
adjustment
of
1­
2
days
is
conservative,
in
which
case
the
estimated
postchlorination
measurement
date
is
between
4­
28­
99
and
4­
29­
99.
Therefore,
the
estimated
hold
time
for
this
sample
is
1
day.

Liberty,
MO,
990443A
This
sample
was
received
at
13:
15
on
4­
27­
99
as
is
indicated
on
the
Biovir
Laboratories
Inc.
Sample
Transfer
Receipt.
Because
this
sample
arrived
late
in
the
day,

EPA
assumed
that
work
began
on
this
sample
on
4­
28­
99.
Thus,
4­
28­
99
was
established
as
the
estimated
pre­
chlorination
measurement
date.
The
date
on
the
total
coliform
study
sheet
for
this
sample
was
4­
30­
99
(
no
date
is
available
for
the
E.
coli
results).
In
light
of
the
above
discussion,

and
considering
that
post­
chlorination
measurements
cannot
precede
pre­
chlorination
measurements,
an
adjustment
of
1­
2
days
is
conservative,
in
which
case
the
estimated
postchlorination
measurement
date
is
between
4­
28­
99
and
4­
29­
99.
Therefore,
the
estimated
hold
time
for
this
sample
is
1
day.

Ames,
IA,
990438A.
This
sample
was
received
at
10:
15
on
4­
27­
99
as
is
indicated
on
the
Biovir
6
Laboratories
Inc.
Sample
Transfer
Receipt.
Because
this
sample
arrived
early
in
the
day,
EPA
assumed
that
work
began
on
this
sample
on
4­
27­
99.
Thus,
4­
27­
99
was
established
as
the
estimated
pre­
chlorination
measurement
date.
The
date
on
the
total
coliform
study
sheet
for
this
sample
was
4­
30­
99
(
no
date
is
available
for
the
E.
coli
results).
In
light
of
the
above
discussion,

and
considering
that
post­
chlorination
measurements
cannot
precede
pre­
chlorination
measurements,
an
adjustment
of
1­
3
days
is
conservative,
in
which
case
the
estimated
postchlorination
measurement
date
is
between
4­
27­
99
and
4­
29­
99.
Therefore,
the
estimated
hold
time
for
this
sample
is
no
more
than
2
days.
