November
23,
2005
To:
Steve
Beaulieu,
Steve
Neubold
Rich
Iovanna
From:
Dennis
Scarnecchia
Subject:
Maximum
annual
reproductive
rates
among
species
groups.

As
a
result
of
our
past
conference
call,
I
was
asked
to
provide
a
professional
opinion
of
the
use
of
the
maximum
annual
reproductive
rate
(
log
alpha)
used
in
a
paper
by
Myers
(
1999.
Can.
J.
Fish.
Aquat.
Sci.
56:
2404­
2419).
Specifically,
as
I
understood
it,
Steve
N.
and
Rich
were
interested
in
an
assessment
of
whether
the
log
alpha
values
representing
the
log
of
the
maximum
reproductive
rate,
in
Myers'
Table
1
(
Page
2412)
for
a
diversity
of
species
could
be
reasonably
attributable
to
other
taxonomically­
related
species
not
listed
in
the
table.
This
would
allow
the
modeling
effort
to
proceed
for
those
species
for
which
data
on
maximum
reproductive
rates
were
not
available.
In
response
to
that
question,
I
have
reviewed
the
paper
on
three
different
occasions,
related
it
to
the
standard
alpha
parameter
in
the
Ricker
model
on
which
it
is
based,
e­
mailed
Dr.
Myers
with
a
couple
of
specific
questions,
and
recently
(
yesterday)
finally
gotten
to
speak
with
him
for
about
30
minutes
to
discuss
the
attributes
and
limitations
of
the
concept
as
outlined
in
the
paper.
Based
on
the
above
activities,
I
offer
the
following
evaluation
and
recommendations:

The
concept
of
the
maximum
annual
reproductive
rate
as
outlined
in
the
paper
seems,
at
least
at
first,
like
a
powerful
and
useful
concept,
and
Myers
certainly
thinks
that
it
is.
I
am
less
sure
of
it,
although
the
paper
is
certainly
a
creative
work
to
generate
discussion.
He
makes
a
strong
point
in
the
paper
that
the
rates
only
vary
between
1
and
7.
The
reasons
for
this
limited
range
of
values
could
be
many,
but
the
most
likely
reason
is
that
there
are
always
environmental
reasons
why
an
organism
can
only
reproduce
so
rapidly,
evan
at
low
population
sizes
when
intraspecific
density­
dependent
effects
are
minimal
or
non­
existent.
For
fish,
these
reasons
may
in
a
very
few
cases
relate
to
fecundity
for
a
few
viviparous
species
or
other
species
with
very
low
reproductive
rates,
but
for
most
species
it
is
not
the
fecundity
that
is
the
problem:
there
are
usually
plenty
of
eggs.
The
problem
is
usually
the
other
mortality
factors
that
act
on
all
those
eggs,
in
the
form
of
predation
on
them
or
the
young
by
other
species,
food
limitations
on
the
young
because
of
competition
with
other
species,
etc.,
that
limit
how
fast
a
population
can
multiply
even
when
there
are
few
of
the
species
out
there.
In
simple
terms,
things
are
usually
tough
out
there
in
nature
for
one
reason
or
another,
and
that
keeps
maximum
reproductive
rates
down
to
7
or
below
in
all
but
unusual
circumstances.
In
the
unusual
circumstances,
such
as
a
quickly
established
invasive
species
with
abundant
food
and
few
or
no
predators/
competitors,
the
maximum
annual
reproductive
rate
may
be
quite
a
bit
higher
for
a
few
years
(
e.
g.
a
population
or
pest
boom)
but
it
will
not
stay
that
way:
something
will
happen
to
reduce
reproductive
success
in
the
future.
In
terms
of
the
Ricker
model,
it
means
that
the
slope
of
the
stock­
recruitment
curve
will
typically
be
only
so
steep,
and
not
too
steep
on
average,
implying
that
the
stock­
recruitment
curves
will
tend
to
fall
within
a
restricted
range
of
alpha
values.
Myers
indicated
to
me,
for
example,
that
the
large
majority
of
stocks
of
sockeye
salmon
tend
to
have
higher
alpha
values
(
about
5)
than
most
stocks
of
pink
salmon
(
about
3),
so
in
his
view
there
are
species
specific
trends.
This
makes
sense
because
the
life­
history
differences,
as
well
as
the
rigors
they
might
encounter,
may
be
more
similar
among
sockeye
stocks
and
more
similar
among
pink
salmon
stocks
than
between
the
two
species.
I
also
asked
him
if
in
his
view
there
was
a
tendency
for
life
history
strategy
to
play
a
role,
such
that,
as
I
had
observed
in
the
Table
1,
the
long­
lived,
late­
maturing
species
tended
to
have
lower
alpha
values
than
the
short­
lived,
early­
maturing
species.
If
so,
then
there
would
be
a
good
rationale
for
using
the
log
alpha
values
for
one
species
for
closely
related
species.
He
said
there
was
some
tendency
for
what
I
thought
I
had
observed
to
be
so,
but
in
his
opinion,
there
was
not
a
terribly
strong
relationship
between
the
two
 
just
a
weak
one.
He
did
not
seem
to
me
to
be
very
sure
about
it.
He
was
also
a
bit
unsure
about
the
quality
of
some
of
his
data
sets
(
e.
g.,
sable
fish
and
others)
that
had
low
log
alpha
values.
He
made
this
point
to
me
as
well
over
the
phone
as
in
the
paper.
To
me
the
question
boils
down
to
an
analysis
of
variance:
what
is
the
variance
in
alpha
values
within
stocks
of
a
species,
among
species,
and
among
families/
classes
or
life
histories
of
fish,
and
what
are
the
causes.
Myers
did
not
relate
any
clear
answers
to
this
question
to
me.
He
did
not
have
any
answers
to
share
on
this.
I
believe
the
problem
is
that
no
one
is
sure
of
the
answer.

Here
are
my
conclusions,
opinions,
and
recommendations.

1.
EPA
modelers
should
keep
in
mind
that
even
though
table
1
has
many
species
listed
and
looks
at
many
time
series,
a
few
species
and
species
groups
are
heavily
overrepresented
(
Clupeidae,
Gadidae,
Salmonidae)
while
most
other
species
are
badly
underrepresented
In
the
under­
represented
situations,
there
are
many
reasons
why
a
particular
alpha
has
its
particular
value
and
why
it
may
not
represent
the
species
well.
One
reason
could
be
that
the
data
used
in
the
calculation
are
poor,
a
situation
repeatedly
invoked
by
Myers
himself
in
the
paper
and
in
my
conversation
with
him.
A
second
reason
could
be
that
the
population
may
be
in
an
uncommonly
good
or
bad
environmental
period
that
would
make
alpha
higher
or
lower
(
respectively)
than
it
would
be
under
average
environmental
conditions.
This
is
because
the
Ricker
model
and
Myers'
model
are
deterministic
and
did
not
considering
environmental
nor
other
(
stochastic)
changes.
In
this
situation,
then,
a
number
of
different
stocks/
populations
of
a
given
species
with
a
fairly
constant
life
history
strategy
among
stocks
may
have
different
alpha
values
depending
on
what
different
environmental
factors
they
encounter.

2.
Based
on
my
experience
with
life
history
strategies
and
stock­
recruitment
curves,
it
is
my
opinion
that
if
only
truly
good
data
sets
were
used
(
they
are
seldom
available)
whose
strengths
and
limitations
were
well
understood
by
the
scientist
(
rarely
the
case!),
there
would
be
some
general
discernable
patterns
to
the
variations
in
log
alpha
values
observed.
Over
time,
on
average,
there
would
be
a
tendency
for
alpha
values
to
cluster
for
similar
species
with
similar
life
history
strategies.
Short
lived,
early
maturing
species
would
tend
to
be
higher
in
alpha
values
than
long­
lived,
late
maturing
species.
I
cannot
see
that
it
would
not
be
so:
to
me,
the
long­
lived,
late
maturing
species
make
up
for
their
low
productivity
by
hanging
in
there
with
a
low
mortality
rate
over
the
long
term.
In
this
situation,
it
would
be
the
life
history
strategy
rather
than
the
taxonomic
group
per
se
that
would
result
in
some
common
range
of
alpha
values,
although
because
life
histories
of
related
taxa
are
often
more
similar
than
not,
their
alphas
would
also
be
more
similar
than
not.
So,
for
example,
I
would
expect
that
for
herrings,
alpha
values
would
clustered
because
there
is
not
too
great
a
range
in
life
histories
among
them.
Maybe
the
same
for
cod,
in
general.
My
view
may
be
overly
simplistic,
but
this
is
how
I
see
it.
For
that
reason,
I
think
that
there
is
validity
in
using
species
with
similar
taxonomy
and
life
histories
as
surrogates
for
each
other
in
the
analysis.
I
would
caution,
however,
that
the
value
you
use,
if
it
is
from
just
one
or
two
stocks
in
one
or
two
situations,
may
not
represent
that
taxon/
strategy
adequately.
I
would
feel
safer
using
data
from
a
species/
strategy
group
with
more
time
series,
and
taking
the
median
from
them
all,
for
example.

I
hope
that
this
is
helpful.
We
can
discuss
at
208:
885­
5981.
I
will
be
in
Friday
after
Thanksgiving
and
all
next
week.
My
apologies
for
the
delay
in
this
response.
Happy
Thanksgiving.

Dennis
