Experimental Studies on the Duration of Life. II. Hereditary differences in Duration of Life in Line-bred Strains of Drosophila


Pearl, R.; Parker, S.L.

American Naturalist 56(643): 174-187

1922


The data presented in this paper appear to demonstrate, with comprehensiveness and accuracy, three broad facts. A. That there exist in a general population of Drosophila melanogaster (or its mutants) genetic differences in respect of duration of life. B. That these genetic differences are capable of isolation, by appropriate selection and inbreeding. C. That within an even moderately inbred line, the genetic differences in duration of life remain constant over periods of at least 10 to 25 or more generations. These facts, based upon the determination experimentally of the duration of life of 3,039 individual flies in 18 experiments, under constant environmental conditions, place this character "duration of life" in the category of genetically definite and workable characters, and indicate that it will just as well repay careful analytical study as the characters more usually dealt with. Furthermore, duration of life is a character of great general biological significance.

EXPERIMENTAL
STUDIES
ON
THE
DURA-
TION
OF
LIFE
II.
HEREDITARY
DIFFERENCES
IN
DURATION
OF
LIFE
IN
LINE
-BRED
STRAINS
OF
DROSOPHILA
PROFESSOR
RAYMOND
PEARL
AND
SYLVIA
L.
PARKER
INTRODUCTION
IT
was
shown
in
the
fi
rst
paper
in
this
series
(27)
2
that
there
was
a
marked
difference
in
mean
duration
of
life,
and
in the
form
of
the
curve,
between
wild
-type
stocks
of
Drosophila
on
the
one
hand
and
the
syntheuic
quin-
tuple
mutation
stock
on
the
other
hand.
It
was
further
made
clear
that,
because
of
the
technique
used
in
the
ex-
perimental
work,
there
could
be
no
doubt
that
the
basis
of
this
difference
must
be
hereditary
and
not
environmental.
Furthermore,
Hyde
(11)
and
Pearl
(6)
have
presented
evidence
for
the
Mendelian
inheritance
of
this
character
duration
of
life.
Given
it
to
be
the
fact,
as
the
just
cited
work
demon-
strates
to
be
the
case,
that
there
are
hereditary
differ-
ences
within
the
same
species
of
Drosophila
in
respect
of
duration
of
life,
the
problem
which
next
presents
itself
is
to
determine
whether
within
a
particular
strain
of
Droso-
phila
hereditary
differences
exist,
and
if
so
what
their
magnitude
may
be,
their
degree
of
permanence,
etc.
In
1
Papers
from
the
Department
of
Biometry
and
Vital
Statistics,
School
of
Hygiene
and
Public
Health,
The
Johns
Hopkins
University,
No.
48.
2
A
word
of
explanation
is
necessary
as
to
the
method
of
handling
biblio-
'
graphic
references
in
this
series
of
papers.
In
the
fi
rst
paper
a
list
of
26
references
numbered
consecutively
from
1
was
appended.
It
is
proposed
not
to
duplicate
references
in
any
subsequent
paper
in
the
same
series.
Conse-
quently
the
fi
rst
itew
bibliographic
citation
in
the
present
paper
is
numbered
27.
When
any
reference
is
made
to
titles
already
cited
in
the
fi
rst
paper
in
the
series,
the
numbers
which
they
bear
in
the
list
appended
to
that
paper
will
be
used.
This
practice
will
be
adhered
to
in
all
subsequent
papers
in
this
series
of
Studies.
174
No.
643]
STUDIES
ON
THE
DURATION
OF
LIFE
175
short
one
wishes
immediately
to
get
a
kind
of
knowledge
for
this
organism
and
character
similar
to
that
which
Johannsen
(28,
29)
got
for
the
size
character
of
beans
from
his
pure
-line
work.
The
fi
rst,
and
in
a
sense
pre-
liminary,
investigations
on
this
problem
will
be
presented
in
this
paper.
Later
in
the
series
we
expect
to
publish
much
more
extended
and
penetrating
evidence
on
the
same
problem.
Some,
however,
must
be
presented
early
in
the
series
in
order
to
make
the
account
of
subsequent
experiments
intelligible.
It
is
obvious
that
in
the
case
of
an
organism
like
Droso-
phila
it
is
impossible
to
have
a
pure
-line
in
the
strict
sense
of
Johannsen.
The
most
that
one
can
do
is
to
have
inbred
lines,
and
the
most
intense
degree
inbreeding
possible
in
the
pr,mises
is
by
brother
X
sister
mating.
The
gen-
eral
plan
of
the
experiments
reported
in
this
paper
can
be
outlined
as
follows
:
1.
Mate
a
virgin
brother
and
sister,
chosen
at
random
each
from
the
same
one
of
the
original
5
foundation
stocks
(cf.
27).
2.
Repeat
this
for
as
many
pairs
as
the
facilities
of
the
laboratory
make
possible.
3.
Test
the
progeny
of
each
mated
pair
separately
for
duration
of
life,
and
form
for
each
group
of
such
progeny
a
life
table.
4.
Each
such
mated
pair
constitutes
the
beginning
of
a
line,
in
which
at
any
time
the
processes
noted
under
para-
graphs
1,
2,
and
3
above
could
be
repeated.
In
this
paper
will
be
reported
the
results
of
one
such
repetition.
The
general
technique
of
the
experimental
work
has
been
fully
described
in
the
fi
rst
paper
of
this
series
and
need
not
be
repeated.
It
should
merely
be
emphasized
again
that
the
environmental
conditions
in
respect
of
food,
housing,
temperature
(25°
C.)
and
atmospheric
con-
ditions
were
identical
for
all
the
fl
ies
in
the
experiments
here
reported.
176
THE
AMERICAN
NATURALIST
[VoL.
LVI
DURATION
OF
LIFE
IN
DIFFERENT
PROGENY
GROUPS
OUT
OF
BROTHER
X
SISTER
MATINGS
The
survivorship
data
frequencies)
for
7
progeny
groups
each
out
of
a
mating
of
brother
X
sister
are
ex-
hibited
in
Table
II.
All
distributions
are
put
on
the
same
basis
of
1,000
fl
ies
at
emergence
from
the
pupal
stage.
The
absolute
numbers
of
fl
ies
involved
in
each
experiment
are
given
at
the
foot
of
each
column.
These
numbers
are
TABLE
I
BROTHER
X
SISTER
MATINGS.
FIRST
TEST
Lines
Original
Stock
(Described
in
(27))
Date
of
Mating
Date
of
Emergence
100
Old
Falmouth
April
8,
1920
April
19
-May
3
101
April
7,
1920
App
it
17
-May
2
201
New
Falmouth
April
7,
1920
April
18
-May
2
202
April
10,
1920
April
20
-April
29
300
Sepia
April
7,
1920
April
17
-May
3
301
April
6,
1920
April
17
-May
3
303
April
8,
1920
April
18
-May
2
TABLE
II
SURVIVORSHIP
DISTRIBUTIONS
OF
PROGENY
OF
BROTHER
X
SISTER
MATINGS.
BOTH
SEXES
TOGETHER
Age
in
Days
Numbers
of
Survivors
up
to
Indicated
Age
in
Lines
No.
100
101
201
202
300
301
303
1,000
1,000
1,000
1,000 1,000
1,000
1,000
6
983
993
1,000
G89
926
870
882
12
937
987
1,000
607
858
727
764
18
891
934
952
492
809
602
702
24
811
901
943
426
G23
441
621
30
743
875
857
344
549
342
522
86
589
855
790
197
383
255
429
42
514
770
600
148
272
205
311
48
406
599
505
148
148
130
261.
54.
240
493
381
49
105
75
168
60
91
29G
219
33
12
31
112
66
29 99
133
1G
6
12
56
72
6
10
0 0
2
0
78
0
7
10
0
84
7
0
90
0
Abs.
No.
of
fl
ies
175
152
105
61
162
161 161
No.
643]
STUDIES
ON
THE
DURATION
OF
LIFE
177
smaller
than
is
desirable,
but
these
experiments
represent
a
relatively
early
stage
of
the
work
before
the
technique
of
getting
maximum
progenies
for
life
table
work
had
been
perfected.
Further
it
must
be
remembered
that
the
individuals
in
any
column
are
the
progeny
of
only
one
single
pair
of
parents.
The
source
of
the
lines
together
with
other
pertinent
data
are
shown
in
Table
I.
1000
''
••••,
N
/00
\
co
\
'
\
/0
1 1
1 1 1 1
1
1
1 1 1,
1
\
I I
12
16
24
30
36
42
48
54
60
66
72
78
54
90
AGE
IN
DAYS
FIG.
1.
Survivorship
(t
i
)
graphs
for
lines
100, 101,
201,
202
and
301.
Five
of
these
distributions
are
shown
graphically
in
Fig.
1,
and
their
biometric
constants
are
given
in
Table
TABLE
III
FREQUENCY
CONSTANTS
FOR
cl„
DISTRIBUTIONS.
FIRST
TEST
Line
No.
Mean
Duration
of
Life
(Days)
Standard
Deviation
(Days)
Coefficient
of
Variation
100
40.45
t
.84
16.38
t
.59
40.49
t
1.68
101
50.02
t
.85
15.51
t
.60
33
1
1.
.0
7
1
1
1
1.
.
5
31
1
201
47.40
t
.99
15.03
t
.70
202
22.04
t
1.57
18.18
t
1.11
82.49
t
7.74
300
31.19
t
.83
15.76
t
.59
50.53
t
2.33
301
25.28
t
.92
17.25
t
.65
68.24
t
3.56
303
32.02
t
1.07
20.04
t
.75
62.59
t
3.14
178
THE
AMERICAN
NATURALIST
[VoL.
LVI
III.
In
calculating
these
constants,
the
absolute
dx
fre-
quencies,
and
not
the
per
mille
frequencies,
were
of
course
used.
From
these
data
it
is
at
once
apparent
that
these
progeny
groups
show
distinct,
and
in
some
cases
decidedly
large,
differences
both
in
mean
duration
of
life
and
in
the
form
of
the
mortality
distributions.
Lines
101
(Old
Fal-
mouth
stock)
and
201
(New
Falmouth
stock)
show
the
longest
mean
duration
of
life,
and
they
are
sensibly
iden-
tical
in
the
form
of
the
life
curve,
having
regard
to
the
errors
of
random
sampling.
The difference
in
the
means
for
these
two
lines
is
2.62
±
1.31
days,
an
obviously
insig-
nificant
difference,
only
2
times
its
probable
error.
Simi-
larly
these
two
lines
do
not
significantly
differ
in
absolute
or
relative
variability,
the
difference
between
the
stand-
ard
deviation
being
.48
±
.92.
Line
100
(Old
Falmouth
stock)
has
a
distinctly
and
sig-
nificantly
lower
mean
duration
of
life
than
101
or
201.
Comparing
it
with
line
101
the
difference
in
the
means
is
9.57
±
1.20
days
or
approximately
8
times
its
probable
error.
The
l
curve
lies
throughout
its
course
below
the
lines
for
101
and
201.
Line
100
is
also
relatively
more
variable
in
duration
of
life
than
101
and
201,
but
largely
because
of
the
difference
in
the
means.
The
individuals
in
line
202
(New
Falmouth
stock)
are
the
shortest
lived
of
any
here
dealt
with,
and
the
shortest
-
lived
wild
-type
strain
we
have
as
yet
isolated.
Its
mean
duration
of
life
is
less
than
half
that
shown
by
lines
101
and
201
and
only
a
little
more
than
half
that
of
line
100.
Line
202
shows
the
highest
relatiVe
variability
in
duration
of
life
of
any
of
the
lines
here
discussed.
It
also
has
the
highest
absolute
variability
with
one
exception
(line
303).
Lines
300,
301
and
303
(Sepia
stock)
are
all
relatively
short-lived
lines.
300
and
303
are
substantially
identi-
cal,
while
301
has
a
lower
mean
approaching
that
of
line
202.
These
sepia
lines
are
also
characterized
by
high
relative
variability.
No.
643]
STUDIES
ON
THE
DURATION
OF
LIFE
179
RESULTS
OF
INBRED
RE
-TESTS
FOR
CONSTANCY
During
the
progress
of
the
experiments
described
in
the
preceding
section
the
offspring
fl
ies
(from
original
brother
X
sister
matings)
in
each
of
the
lines,
whose
dura-
tion
of
life
was
being
tested,
were
allowed
to
mate
at
random
in
their
bottles,
and
their
progeny
removed
to
form
stocks
of
the
several
lines.
These
stocks
were
al-
lowed
to
reproduce
in
stock
bottles,
all
matings
being
therefore
random
within
the
line,
for
a
period
of
about
7
months
(cf.
Table
IV).
At
the
end
of
that
time
it
was
de-
cided
to
make
a
re
-test
of
each
line
to
see
how
it
was
then
behaving
relative
to
duration
of
life.
There
was
then
made,
at
dates
indicated
in
Table
IV,
a
random
selection
from
each
line
stock
bottle
from
which
a
brother
and
sister
pair
was
bred,
and
these
two
individuals
were
mated
to
get
a
set
of
progeny
on
which
to
carry
out
a
sec-
ond
set
of
life
duration
experiments.
The
necessary
facts
as
to
line
numbers
and
dates
on
this
re
-test
are
given
in
Table
IV.
TABLE
IV
BROTHER
X
SISTER
MATINGS.
SECOND
TEST
Line
from
which
Second
Selection
of
Brother
and
Sister
Was
Made
Number
of
Line
of
Progeny
of
Second
Brother
X
Sister
Mating
Date
of
Original
Brother
X
Sister
Mating
Date
of
Second
Brother
X
Sister
Mating
100
104
April
8,
1920
November
6,
1920
101
107
April
7,
1920
October
14.
1920
201
207
April
7,
1920
October
18,
1920
202
208
April
10,
1920
October
14,
1920
300
304
April
7,
1920
November
6,
1920
301
April
6,
1920
October
14,
1920
303
309
April
6,
1920
October
14,
1920
The
survivorship
distributions
of
the
progeny
groups
of
this
second
brother
X
sister
mating
are
given
in
Table
V,
and
the
biometric
constants
calculated
from
the
ob-
served
6/,
distributions
in
Table
VI.
These
tables
are
for
comparison
with
Tables
II
and
III
above.
180
THE
AMERICAN
NATURALIST
[VoL.
LVI
TABLE
V
SURVIVORSHIP
DISTRIBUTIONS
OF
PROGENY
OF
SECOND
BROTHER
X
SISTER
MATINGS.
BOTH
SEXES
TOGETHER
Age
in
Days
Numbers
of
Survivors
up
to
Indicated
Age
in
Lines
No.
104
107
207
208
304
307
309
1
1,000
1,000
1,000
1,000
1,000
1,000
1,000
6
997
1,000
973
833
1,000
862
1,000
12
923
950
926
738
870
700
978
1
8
871
926
819
643
870
623
911
24
713
917
792
595
674
469
700
30
629
901
785
500
478
392
489
36
552
860
711
286
435
285
456
42
469
777
644
167
261
177
267
48
395
686
530
0
152
92
89
54
304
595
430
109
46
67
60
178
488
255
0
23
44
66
66
264
141
8
0
72
0
83
20
8
78
8
20
8
84
0
7
0
90
-
-
0
Abs.
No.
of
fl
ies
286
121
149
42
46
130
90
TABLE
VI
FREQUENCY
CONSTANTS
FOR
d
DISTRIBUTIONS.
SECOND
INBRED
TEST
Line
No.
Mean
Duration
of
Life
(Days)
Standard
Deviation
(Days)
Coefficient
of
Variation
104
39.59
t
.74
18.63
t
.53
47.06
t
1.62
107
53.74
t
1.07
17.40
t
.75
32.38
t
1.54
207
45.34
t
1.10
19.97
t
.78
44.04
t
2.03
208
25.65
t
1.53
14.68
t
1.08
57.23
t
5.42
304
32.09
1.43
14.43
t
1.01
44.97
t
3.75
307
25.22
t
.99
16.70
t
.70
66.22
t
3.79
309
33.00
t
.91
12.84
t
.65
38.91
t
2.23
The
purpose
of
this
second
test
was,
of
course,
to
see
to
what
extent
duration
of
life
was
holding
constant
in
the
line.
During
the
period
between
the
fi
rst
and
second
test
the
stocks
of
the
several
lines
had
been
subjected
to
vary-
ing
environmental
influences,
in
particular
in
relation
to
temperature,
the.
stock
bottles
having
been
kept
at
room
temperature,
which
varied
rather
extensively.
Did
the
lines
after
7
months
have
the
same
characteristic
life
curves
that
they
exhibited
on
the
fi
rst
test?
Allowing
12
No.
643]
STUDIES
ON
THE
DURATION
OF
LIFE
181
days
from
generation
to
generation
in
the
case
of
flies
re-
producing
freely
at
random
in
stock
bottles,
the
interval
elapsing
between
the
fi
rst
and
second
tests
would
cover
roughly
almost
18
generations.
This
is
a
long
period
and
affords
abundant
opportunity
for
change
in
the
aver-
age
genetic
constitution
of
the
population.
1000
100
/0
FIG.
2.
<//v
e
2/,
yZ,
,
-
‘/Ne
00
6
12
18
24
30
36
42
48 54
60
66
72
78
84
90
AGE
IN
DAYS
Comparing
the
i,
lines
of
the
fi
rst
and
second
inbred
tests.
of
lines
101,
100
and
301.
An
examination
of
Tables
V
and
VI
and
Fig.
2
shows
at
once,
in
a
general
way,
that
the
characteristic
features
of
the
several
lines
in
respect
of
duration
of
life
did
in
fact
hold
remarkably
constant
during
this
period.
A
more
precise
comparison
of
the
means
is
made
in
Table
VII.
There
can
be
no
question
of
the
substantial
constancy
of
these
lines,
over
the
period
covered
in
the
tests
in
re-
spect
of
duration
of
life.
The
L
curves
run
well
together
till
the
upper
end
of
life
is
reached,
where,
because
of
the
small
numbers
involved,
there
is
some
irregularity.
In
no
case
is
the
difference
between
two
comparable
means,
as
-shown
in
Table
VII,
as
much
even
as
three
times
its
prob-
182
THE
AMERICAN
NATURALIST
[VoL.
LVI
able
error,
nor
is
there
any
certainly
significant
change
in
variability
having
regard
to
the
probable
errors
of
the
differences
involved.
TABLE
VII
DIFFERENCES
IN
MEAN
DURATION
OF
LIFE
BETWEEN
THE
FIRST
AND
SECOND
INBRED
TESTS
OF
THE
SEVERAL
LINES
Corresponding
Lines
(Mean
of
Second
Test
Minus
Mean
of
First)
104-100
Difference
of
Means
(Days)
.86
+
1.12
Diff.
P.
E.
Diff.
.77
107-101
-1-
3.72
±
1.37
2.72
207-201
2.06
±
1.48
1.39
208-202
-1-
3.61
H-
2.19
1.65
304-300
-1-
.90
±
1.65
.54
307-301
-1-
.06
-I
-
1.35
.04
309-303
-1-
.98
I-
-
1.40
.70
RESULTS
OF
MASS
CULTURE
RE
-TESTS
FOR
CONSTANCY
The
point
may
well
be
made
that
in
the
re
-tests
of
the
lines
described
in
the
preceding
section
an
additional
ele-
ment
is
introduced
in
the
fact
that
the
fl
ies
for
the
re
-test
were
the
progeny
of
a
second
brother
X
sister
mating.
What
one
wishes
to
know
is:
what
degree
of
constancy
in
duration
of
life
is
exhibited
by
the
general
stocks
in
each
line,
mating
purely
at
random,
after
the
initial
se-
lection
and
inbreeding?
We
wish
now
to
present
some
data
on
this
point.
Table
VIII
gives
the
biometric
con-
stants
for
this
material.
Mass
culture
re
-tests
have
been
made
on
two
of
the
original
lines,
100
and
101.
These
mass
culture
re
-tests
were
made
in
two
ways
as
follows:
(a)
From
the
stock
bottles
of
the
line
to
be
tested
a
large
sample
of
progeny
was
taken
at
random
each
day
as
the
fl
ies
emerged
from
the
pupal
stage,
and
these
progeny
fl
ies
were
put
in
small
bottles
for
a
duration
of
life
experiment
in
the
usual
way
described
in
(27).
(b)
From
the
stock
bottles
of
a
particular
line
to
be
tested
a
number
of
virgin
flies
(usually
8
to
10
of
each
sex)
were
taken
at
random
immediately
upon
emergence,
and
mated
as
a
group
in
a
mating
bottle.
The
progeny
from
this
sample
was
then
remOved,
upon
emergence,
to
small
bottles
and
a
regular
duration
of
life
test
carried
as
described
in
(27).
0
ca
TABLE
VIII
FREQUENCY
CONSTANTS
FOR
MASSI
CULTURE
RE
-TESTS.
ORIGINAL
LINES
Line
No.
Test
Dates
of
Emergence
Numbers
of
Flies
Mean
(Days)
Standard
Devia-
tion
(Days)
Coefficient
of
Variation
100
Mass
Culture
1920
Sept.
25
-Oct.
21
433
33.05
t
.76
23.48
t
.54
71.05
±
2.31
Original
brother
X
sister
1920
Apr.
19
-May
3
175
40.45
±
.84
16.38
±
.59
40.49
±
1.68
Difference
5
mo.
12
days
- -
7.40
t
1.13
+
7.10
±
.80
+30.56
±
2.68
101
Mass
Culture
A
1920
Sept.
25
-Oct.
21
473
53.09
±
.70
22.53
±
.49
42.44
t
1.09
Mass
Culture
B
1921
Mar.
18
-Apr.
4
,
124
48.53
±
1.02
16.76
t
.72
34.54
t
1.65
Original
brother
X
sister
1920
Apr.
17
-May
2
152
50.02
±
.85
15.51
±
.60
31.01
±
1.31
Difference
OA
5
mo.
13
days
- +
3.07
t
1.10
+
7.02
t
.77
+11.43
±
1.70
Difference
OB
10
mo.
29
days
-
-
1.49
t
1.33
+
1.25
±
.94
+
3.53
t
2.11
STUDIES
ON
THE
DURATION
OF
LIFE
00
184
THE
AMERICAN
NATURALIST
[VoL.
LVI
It
is
at
once
apparent
that
the
mass
re
-tests
on
line
101
gave
extremely
satisfactory
results
as
to
constancy
of
duration
of
life
in
the
line,
after
intervals
of
approxi-
mately
5
and
11
months.
The
mean
value
for
either
the
A
or
the
B
test
does
not
significantly
differ,
having
regard
to
its
probable
error,
from
the
mean
shown
on
the
original
test
at
the
start
of
the
line.
The
mean
of
the
A
mass
re-
test
almost
exactly
agrees
with
that
of
the
second
inbred
test
of
the
same
line,
as
given
in
Table
VI.
In
the
case
of
line
100,
the
mass
re
-test
after
5
months
approximately
does
not
give
such
close
agreement.
The
mean
is
significantly
lower,
the
difference
being
6.6
times
its
probable
error.
No
explanation
of
this
result
is,
as
yet,
forthcoming,
but
it
probably
means
no
more
than
lack
of
genetic
purity
in
the
line.
It
is,
however,
interesting
to
note
that
the
sense
of
the
change
is
in
the
same
direction
as
that
in
which
line
100
in
general
differs
.,from
line
101,
which
we
regard
as
our
most
typical
wild
-type
line
in
re-
spect
of
duration
of
life.
That
is,
line
100
is,
as
com-
pared
with
101,
a
shorter
-lived
line.
Its
mass
culture
re
-test
is
still
shorter
lived.
The
variability
in
respect
of
duration
of
life,
whether
measured
in
absolute
or
relative
terms,
is
uniformly
higher
and
in
two
cases
out
of
the
three
by
a
significant
amount
in
the
mass
culture
than
in
the
original
inbred
tests.
This
is,
of
course,
exactly
what
would
be
expected
on
general
genetic
grounds.
One
brother
X
sister
mating,
as
has
been
shown
by
Pearl
(30),
Jennings
(31)
and
others,
reduces
the
heterozygosis
in
the
strain
by
only
50
per
cent.
It
is
interesting
to
note,
in
connection
with
tie
explanation
suggested
above
for
the
difference
in
the
means
in
the
case
of
line
100,
that
the
variability
in
the
mass
re
-test
on
that
line
is
very
much
higher
than
in
the
original
inbred
test.
A
mass
re
-test
was
carried
out
on
two
of
the
lines
from
the
second
brother
X
sister
matings.
The
results
from
these
experiments
are
presented
in
Table
IX.
z
rF.
TABLE
IX
FREQUENCY
CONSTANTS
FOR
MASS
CULTURE
RE
-TESTS.
TWICE
INBRED
LINE
Line
No.
Test
Mass
Culture
A
Date
of
Emergence
Numbers
of
Flies
Mean
(Days)
Standard
Devia-
tion
(Days)
Coefficient
of
Variation
107
1921
Apr.
19
-Apr.
22
1,338
49.74
±
.25
13.69
±
.18
27.52
±
.54
107
Second
brother
X
sister
S
1920
Oct.
25
-Nov.
4
121
53.74
±
1.07
17.40
±
.75
32.38
±
1.54
101
Original
brother
X
sister
0
1920
Apr.
17-Ma.7,
2
152
50.02
±
.85
15.51
.60
31.01
±
1.31
Difference
SA
6
mos.
9
days
-
4.00
±
1.10
-
3.71
±
.77
-
4.86
±
1.63
Difference
OA
11
mos.
27
days
-
.28
±
.88
-
1.82
.96
-3.49
1.42
309
Mass
Culture
A
1921
May
18-25
468
34.04
.34
11.02
.24
32.38
±
.79
309
Second
brother
X
sister
S
1920
Oct.
25
-Nov.
1
'
90
33.00
.91
12.81
.65
38.91
±
2.23
303
Original
brother
X
sister.
.
. . . . .
1920
Apr.
18
-May
2
161
32.02
±
1.07
20.04
±
.75
62.59
3.14
Difference
SA
6
mos.
28
days
+
1.04
±
.97
-
1.82
±
.69
-
6.53
±
2.37
Difference
OA
13
mos.
1
(lay
+
2.02
1.12
-
9.02
±
.79
-30.21
±
3.24
3
-
317
WO
ivouvuria
WILL
NO
SWIUILLS
00
186
THE
AMERICAN
NATURALIST
[VoL.
LVI
The
substantial
constancy
of
line
101,
in
both
mass
and
inbred
tests,
is
evident.
In
respect
of
variability
the
line
behaved
somewhat
like
303
discussed
below.
In
line
303
again
the
constancy
of
the
line
in
respect
of
mean
duration
of
life
is
as
definite
as
could
be
expected.
Over
periods
of
approximately
7
and
13
months,
the
mean
duration
of
life
has
not
sensibly
changed,
having
regard
to
the
probable
error
involved.
The
results
respecting
variability
are
somewhat
anomalous.
Both
the
second
inbred
and
the
mass
re
-test
show
variability
of
a
dis-
tinctly
lower
order
than
was
exhibited
by
the
progeny
of
the
original
brother
X
sister
mating.
It
seems
probable
that
the
original
test
by
accident
gave
a
variability
re-
sult
higher
than
was
really
characteristic
of
the
line.
But
the
mass
culture
re
-test
exhibits
a
lower
variability,
not
certainly
significant,
to
be,
sure,
than
the
fi
rst
test
on
line
309.
Of
course
it
is
to
be
expected
that
with
continued
brother
X
sister
mating
the
variability
of
mass
cultures
from
the
line
would
come
nearer
and
nearer
to
that
of
a
further
inbred
lot
of
progeny
from
the
same
line.
Prob-
ably
the
results
of
Table
IX
are
an
expression
of
the
realization
of
such
expectation,
obscured
by
the
fact
that
the
numbers
are
small
and
the
errors
of
sampling
conse-
quently
relatively
large.
DISCUSSION
AND
SUMMARY
The
data
presented
in
this
paper
appear
to
demon-
strate,
with
comprehensiveness
and
accuracy,
three
broad
facts.
A.
That
there
exist
in
a
general
population
of
Droso-
phila
melanogaster
(or
its
mutants)
genetic
differences
in
respect
of
duration
of
life.
B.
That
these
genetic
differences
are
capable
of
isola-
tion,
by
appropriate
selection
and
inbreeding.
C.
That
within
an
even
moderately
inbred
line,
the
gen-
etic
differences
in
duration
of
life
remain
constant
over
periods
of
at
least
10
to
25
or
more
generations.
No.
643]
STUDIES
ON
THE
DURATION
OF
LIFE
187
These
facts,
based
upon
the
determination
experi-
mentally
of
the
duration
of
life
of
3,039
individual
flies
in
18
experiments,
under
constant
environmental
conditions,
place
this
character
"
duration
of
life
"
in
the
category
of
genetically
definite
and
workable
characters,
and
indi-
cate
that
it
will
just
as
well
repay
careful
analytical
study
as
the
characters
more
usually
dealt
with.
Furthermore,
duration
of
life
is
a
character
of
great
general
biological
significance.
LITERATURE
CITED
27.
Pearl,
R.
and
Parker,
S.
L.
Experimental
Studies
on
the
Duration
of
Life.
I.
Introductory
Discussion
of
the
Duration
•of
Life
in
Droso-
phila.
AMERICAN
NATURALIST,
Vol.
55,
pp.
481-509,
1921.
28.
Johannsen,
W.
Ueber
Erblichkeit
in
Populationen
and
in
reinen
Linien.
Jena
(Fisher),
1903.
29.
Johannsen,
W.
Elemente
der
exakten
Erblichkeitslehre,
3d
Edit.,
1913.
30.
Pearl,
R.
On
the
Results
of
Inbreeding
Mendelian
Population;
a
Correc-
tion
and
Extension
of
Previous
Conclusions.
AMERICAN
NATURALIST,
Vol.
48,
pp.
57,
62,
1914.
31.
Jennings,
H.
S.
Formulm
for
the
Results
of
Inbreeding.
AMERICAN
NATURALIST,
Vol.
48;
pp.
693-696,
1914.