Experimental studies on the duration of life. VII. The Mendelian inheritance of duration of life in crosses of wild type and quintuple stocks of Drosophila melanogaster


Pearl, R.; Parker, S.L.; Gonzalez, B.M.

American Naturalist 57: 153-192

1923


In this study, involving the determination of the duration of life of 5,415 individual flies, a cross was made between a long-lived stock of Drosophila (Old Falmouth, wild type) and a short-lived stock (Quintuple). In the F1 generations the progeny were somewhat longer lived than either parent stock. In F2 there was a definite and clear-cut segregation in respect of duration of life, long-lived and short-lived groups reappearing, with virtually identical mean duration of life to those of the original parent stocks. In its genetic behavior duration of life was found to be completely and invariably associated with certain morphological characteristics of the organism, in the sense that no vestigial-winged fly has ever been found, in the entire experience of the laboratory, to be long-lived, and no group of normal-winged flies has ever been found to have a life curve even approaching in form that which is characteristic of vestigial-winged flies. The probable meaning of these results, in relation to genetic phenomena in general, is discussed.

EXPERIMENTAL
STUDIES
ON
THE
DURATION
OF
LIFE
VII.
TUE
MENDELIAN'
INHERITANCE
OF
DURATION
OF
LIFE
IN
CROSSES
OF
Wino
TYPE
AND
QUINTUPLE
STOCKS
OF
DROSOPHILA
MELANOGASTER
I
PROFESSOR
RAYMOND
PEARL,
SYLVIA
L.
PARKER
AND
R.
M.
GONZALEZ
INTRODUCTION
THERE
is
now
abundant
evidence
from
at
least
two
widely
different
organisms,
Drosophila
and
man,
that
if
duration
of
life
be
measured
accurately
for
a
con-
siderable
number
of
individuals,
and
these
measure-
ments
be
then
treated
as
attributes
or
characters
of
the
individuals
which
gave
rise
to
them,
the
results
over
a
series
of
generations
are
in
fact
precisely
as
they
would
be
expected
to
be
if
duration
of
life
were
an
inherited
character,
like
stature
or
eye
color
(of.
Pearl,
R.
(52),
for
a
review
of
the
literature
on
which
this
statement
is
based).
In
the
case
of
mall
the
evidence
is
in
the
main
statistical
in
character
(but
see
Pearl,
R.
(53))
and
gives
directly
no
inkling
as
to
the
mechanism
of
the
in-
heritance.
In
Drosophila
Hyde
(11)
showed
that
the
character
appeared
to
be
inherited
in
a
Mendelian
manner,
and.
Pearl
and
Parker
(32)
have
demonstrated
the
constancy
of
definite
degrees
of
mean
longevity
in
inbred
strains
or
lines
of
the
same
organism.
There
is
needed,
however,
much
more
intensive
investigation
of
the
exact
mode
and
mechanism
of
inheritance
of
this
character
in
Drosophila,
in
spite
of
the
excellent
pioneer
work
of
Hyde,
which
we
have
been
able
to
confirm
in
every
essential
particular.
This
and
the
next
follow-
ing
paper
in
this
series
will
be
devoted
to
the
discussion
of
results
which
have
so
far
been
obtained
in
this
fi
eld
Papas
from
the
Department
of
Biometry
and
Vital
Statistics,
School
of
Hygiene
and
Public
Health,
Johns
Hopkins
University.
No.
84.
153
.154
THE
AMERICAN
NATURALIST
[Voi.
LVII
in
the
work
of
this
laboratory.
The
present
paper
will
give
the
analysis
of
data
resulting
from
extensive
ex-
periments
in
crossing
wild
type
flies
of
our
Old
Fal-
mouth
stock
(cf.
Pearl
and
Parker
(27,
p.
486))
with
the
short-lived
Quintuple
stock.
Complete
life
tables
for
wild
type
fl
ies,
of
which
the
Old
Falmouth
stock
is
representative,
and
for
Quintuple
fl
ies,
have
been
presented
in
the
fi
rst
of
these
Studies.
The
wide
difference
in
the
two
stocks
crossed
in
the
present
experiments
is
sufficiently
indicated
by
the
fact
that
for
wild
type
males
the
expectation
of
life
(e.)
at
emergence
is
41.0
days
and
for
Quintuple
males
is
only
14.2
days.
For
females
the
corresponding
e,
fi
gures
are
38.8
and
15.8
days,
respectively.
We
are
thus
dealing
with
well-defined,
indeed
widely
different,
grades
or
de-
grees
of
the
character
in
the
two
stocks
in
these
experi-
ments.
MATERIAL
AND
METHODS
The
data
on
duration
of
li
fe
of
the
parental
stocks
used
in
these
experiments
were
obtained
from
the
prog-
eny
of
4
individual
(brother
and
sister)
matings
in
each
of
the
two
stocks
Old
Falmouth
and
Quintuple,
made
on
April
7-9,
1920.
The
offspring
fl
ies
emerged
between
April
17
and
May
5,
1920.
The
original
parents
were
transferred
to
new
mating
bottles
8
days
after
the
fi
rst
bottle,
in
each
case,
was
started.
The
F,
flies
were
derived
from
one
mass
mating
of
each
of
the
possible
reciprocal
crosses.
These
matings
were
of
6
pairs
of
fl
ies
each,
taken
at
random
from
the
stock
bottles
of
the
two
sorts
used.
The
matings
were
made
April
20,
1920,
and
the
fl
ies
emerged
April
30
May
14.
The
parents
were
transferred
to
new
breed-
ing
bottles
8
days
after
the
first
bottles
were
started.
The
matings
to
produce
F.
and
F„
generations
are
designated
throughout
this
paper
on
the
standard
plan
described
some
years
ago
by
Pearl
(54).
In
order
to
save
reference
to
that
paper
Table
I
is
reproduced
from
No.
6491
STUDIES
ON
THE
DURATION
OF
LIFE
155
it
here.
In
the
present
case
X
=
Quintuple,
and
Y
=
Old
Falmouth.
TABLE
I
MkrucGs
TO
PRODUCE
I',
n
6
B
C
Number
Indi-
of
visuals
Muting
Mated
A
X
X
AXY
A
X
P
A
X
7,
A
X
A
AXB
AXC
A
X
D
A
X
/C
AXF
A
X
G
B
X
X
13
X
Y
BXP
10
40
11
33
25
35
61
0:3
65
14
10
44
P
8
Inch
F2.
-
Nm
u
ber
h
F.2
id
,
of
0
,
ii
d
n
t
t
in
ec
t
Mated
Mating
visuals
E
F
Number
of
Mating
B
X
Z
46
C
X
/1
1
51
/1
X
13
113
C
X6"
53
BXC
37
D
X
X
22
B
X
D
29
D
X
Y
24
13
x
A
'
05
D
x
P
52
B
X
F
D
X
Z
54
B
X
G
69
D
X
1)
17
C
X
X
I8
D
X
E
4:1
CXY
20
D
X
F
31
C
X
P
48
D
X
G
27
C
X
Z
50
E
X
X
20
CXC
15
11
X
Y
28
C
X
D
at)
E
X
P
56
C
XD
41
//
X
Z
58
1
C
Indi-
viduals
Mn-
Mated
-ring
E
X
11
10
FXF
PI
X
G
47
F
x
X
30
32
F
X
P
00
1
7
X
Z
02
10
X
P'
91
X
G
40
GXX
31
GXY
36
0
X
P
01
GXZ
00
G
X
0
23
Table
II
gives
the
data
regarding
the
matings
to
pro
duce
the
F
2
and
F,
generations.
In
all
the
experiments
recorded
in
this
paper
the
only
one
of
the
fi
ve
somatic
mutant
characters
carried
in
the
Quintuple
stock
to
which
any
attention
was
paid
was
vestigial
wing.
Individuals
were
not
recorded
at
all
in
TABLE
TI
Type
of
Type.
of
Pr
i
o
g
l
ee
n
nY
Bating
DATA
ON
mating
MATINUS
TO
PRODUCT.:
D,
AND
D
m
a
att
e
m
o
g
f
MIN,
of
Parents
:
(all
1920)
v.,
N
Par
Mated
of
Date
of
EnegIBof
enge
1
rogeny
Ne
Progeny
Wild
Type
General
l'ype
of
Mating
!
nen
Date
Emergence
of
Parents
.
01
OTh
_
Ves-
tigied
cP
9
e
x
FL
F
(7)
72
I
z
'
5/10-18
.1
15
10
B
(5)
73
5,11
P
4
1
5/1
5/1
5/10-18
50
7
ns
1'
1
X
(7)
74
5/4
Z
6
IS
5/4
3/4
5/13-21
'
109
22
-t
Ng
X
B
(5)
75
5/4
1' 1'
6 6
Bel
5/4
!
5/13-21
140
56
E.
-
7.
-
;
B
X
F1
D
(3)
76
5/20
P
5/29-0/0
115
10
B
X
10
I)
(3)
77
5/20
Z
1'
2 2
5/7
5/8
98
34
ny
639
148
g
-
g
F,
X
Quintuple
(2)
78
5/20
(Bin.
1
I
5/20
503
5/26-6/4
04
71
yn
B
X
Quintuple
(2)
79
5/20
Quin.
4 4
5/7
.5/20
5/30-6/6
24
44
6:
Pr
X
Quintuple
A
(0)
82
5/20
Quin.
1
1
5/20
5/8
5/29-1155
75
70
t
r
,
IL
X
Quintuple
\
(6p
84
30
Quin.
4
4
i
.
I
gniS
5/20
!
54
30
9
217
266.
B
X
Old
, , , , , ,
a
(8)
SO
5128-0/5
142
B
X
Old
Put
(8)
5/20
91116
175
IL
X
Old
Pal
E
(1)
84
5/20
1
1
5)20
5/S
5/28
6/5
1.63
Pi
X
Old
Pal
(4)
1'
O.1
4
0
5/6
5/20
902
Ix!
1
682
XF/
Vyet.
54
86
0/1
3
3
5/29
6/9-17
76
121
B
X
F
Vest.
52
88
0/1
rhe
3
3
5/10
6/00-18
95
107
Fr
d'
X
Ix
Vest.
9
54
90
077
3 3
5/9
5
/80
2.9
6/9
-17
67
76
52
92
6/1
3
5/10
5/29
6/9
-17
63
119
2
301
423
F,
5
5
X
F,
Wild
9
54
87
1
1
3
D7€
6/10-15
155
5n
Fr
ci
e
X
B2
Wild
9
92
39
5/1
P
Dm
3
5/9
5130
6/9
-17
175
0
Fr
cr
X
1s2
Wild
9
54
91
6/1
Dn
a
0/10-18
103
2
Pa
ce
X
F2
Wild
9
I
52
I
93
1
6/1
Dn
3
5
/
9
5/28
049
-17
152
7
615
61
No.
649]
STUDIES
ON
THE
DURATION
OF
LIFE
157
respect
of
the
other
four,
purple,
speck,
arc
and
black.
This
was
deliberate.
We
desired
to
deal
only
with
that
one
of
the
fi
ve
mutations
which
earlier
experiments
had
shown
to
be
definitely
and
markedly
related
to
duration
of
life,
and
which
at
the
same
time
easily
identified
the
flies
morphologically.
In
the
next
paper
in
this
series
each
of
the
5
mutations
in
Quintuple
will
be
dealt
with
separately
in
relation
to
duration
of
life.
The
experi
ments
here
recorded
were
all
carried
out
in
accordance
with
the
standard
procedure
described
in
the
earlier
papers
in
this
series.
All
determinations
of
duration
of
life
were
made
in incubators
at
25
0
C.
The
food
and
other
environmental
conditions
were
as
constant
and
identical
for
all
groups
as
it
was
possible
to
make
them.
In
tabulating
and
analyzing
the
results
both
sexes
are
treated
together.
TABLE
III
DZATI/
AND
SIIRTIVORSMP
DISTIODUTIONS
FOR
P
AND
r
i
GENERATIONs
Old
ealmou
Age
in
Days
Old
82nininPle
Falmouth'
<X
X
ch
X
Old
'Quintuple
Falmouth
9 9
(A
ating
(Mating
Type
(1))
Type
(0))
II
VI'
I
cl
'
IZ
di 4'
(I.'
2J
I
da'
y
1-
5
19
1,000
1
1,000
2
1
1,000
3
1,000
6-11
14
9701
0
993
6.
094
6
993
12-17
21
948
2
003
,I
077
7
082
18-24
27
915
7
980
4
062
11
968
24-29
38
872
12
934
22
951'
34
946
30-3
63
812
6
854
16
887
22
877
36-41
or
712
12
815
16
841
28
833
42-47
00
609
0
735
(11
795
70
777
48-.3
84
467
18
675
43
618
61
636
34-50
107
334
37
556
01
494
96
513
00-65
74
105
17
311
42
318
69
316
66-71
27
51
14
132
41
197
55
177
78-87
7
14
6
40
24
78
30
06
78-8
0
3-
0
3
9
3
84-89
3
- - -
0
-
90-95
0
Ago
Quintuple
Days
dis
1-
2
10
1,000
3-
5
27
053
6-
8
17
886
9-11
71
760
12-14
72
592
15-17
66
413
18-20
43
240
21-23'
18
142
24-26
19
97
27-20
10
50
0-32
6
25
3-35
3
10
36-38
1
2
0
Totals
1032
—151
-
346
-
407
-
402
-
FIG.
1.
Observed
158
THE
AMERICAN
NATURALIST
[Ver..
LVII
PARENTAL
AND
F,
GRNER.ATIONS
The
observed
death
distributions
and
the
survivor-
ship
distributions
calculated
therefrom,
for
(a)
the
two
parental
stocks,
(b)
the
F,
individuals
from
reciprocal
crosses,
and
(c)
the
combined
F
1
's
are
given
in
Table
III.
In
this
and
all
similar
tables
throughout
the
paper,
the
fi
gures
entered
in
the
di
columns
are
the
observed
number
of
fl
ies
dying
within
the
age
interval
indicated
in
the
age
column,
and
the
figures
in
the
11
column
are
the
numbers
of
survivors
out
of
1,000
starting
imaginal
life
together,
at
the
beginning
of
each
age
interval.
The
survivorship
distributions
of
Table
III
are
shown
graphically,
on
an
arithlog
grid,
in
Fig.
1.
(coo
............
lot
0
6
/d
0
35
42
48
34
E0
el.
/
AGE
IN
DAYS
or
ship
distributions
for
parental
(Old
Falmouth
and
Qnintuple)
and
F
r
.
No.
649]
STUDIES
OW
THE
DURATION
0b'
LIFE
159
The
biometric
constants
calculated
from
the
d.'
dis-
tributions
of
Table
TIT
are
presented
in.
Table
TV.
Ifiromv.
c
Co
Mating
TYPe
Mean
Age
Standard
Cess.ificient
at
Death
Deviation
of
Variation
Parer.
1.Las:south
11.20
A
1
shies
10.57
}
3l
37.44
Quintuple
14.08
-
.23
days
6.93-4.10
-10.23±1.43
o'
X
Ohl
Fal.
P
(0)
Al
.73
-
.57
days
15.71
LL
.-10'
1
30.37
.83
P4
Old
Pal.
d
x
Quint.
9
(1)
51.12
LL
.54
days
!
15.20
±.59
29.903120
All
F
(0)+(1)
.51.50+.47
clays
:
15.553.33
30.23+
.70
TABLE
TV
TS
FOR
DURATION
OP
LIFE
OF
PARENTAL
AND
F,
PLIES
From
the
data
so
far
presented
we
note
the
following
points:
1.
In
these
experiments
we
are
crossing
two
strains
of
Drosophila
of
widely
different
average
duration
of
life
and
forms
of
life
curve.
The
wild
type
Old
Fal-
mouth
parent
stock
has
a
mean
duration
of
life
more
than
three
times
as
great
as
that
of
the
Quintuple
par-
ent
stock.
This
difference
is
so
great,
and
the
two
stocks
are
so
constant
in
the
maintenance
of
their
char-
acteristic
mean
longevities,
that
there
can
be
no
question
that
we
are
crossing
two
distinct
biological
entities
per-
fectly
discontinuous
in
respect
of
the
character
dealt
with.
The
two
stocks
differ
fundamentally
not
only
in
respect
of
mean
duration
of
life
but
also
in
respect
of
variation
in
the
time
of
dying.
2.
The
F,
generation
progeny
from
the
cross
are
structurally
li
ke
the
wild
type
parent.
That
is
to
say,
the
F,
fl
ies
are
to
a
first
degree
of
approximation
indis-
tinguishable
in
their
morphology
from
the
wild
type
parent.
The
morphological
mutations
carried
by
the
Quintuple
fl
ies
behave
as
Mendelian
recessives
to
their
wild
type
allelomorphs.
Also
the
duration
of
life
in
the
F,
flies
is
essentially
similar
to
that
established
by
their
wild
type
(Old
Falmouth)
parent.
They
are
long-lived
fl
ies.
In
actual
fact
the
Tys
have
a
somewhat
longer
mean
duration
of
life
than
the
long-lived,
wild
type
par-
160
THE
AMERICAN
NATURALIST
\m..
tV1I
ent.
Taking
all
F
i
's
together
the
difference
in
the
mean
is
51.55-44.26
=
7.29
4-
.64
days.
This
is
a
certainly
significant
difference
amounting
to
11.4
times
its
prob-
able
error.
The
W
lines
in
Fig.
1
show
plainly
the
su-
perior
longevity
of
the
F
i
's
over
the
longer
lived
parent.
This
phenomenon
is
objectively
similar
to
the
enhanced
vigor
observed
in
the
first
generation
plants
in
Crosses
of
maize,
as
described
by
East
and
Jones
(55),
Shull
(56)
and
others.
It
probably
has
the
same
explanation.
3.
There
is
no
significant
difference
in
respect
of
dura-
tion
of
life
between
the
reciprocal
crosses.
The
mean
duration
of
life
and
the
variability
is
identical
whether
the
cross
-bred
individuals;
had
a
wild
type
father
or
mother.
This
is
to
be
expected
for
the
structural
char-
acters
of
the
fl
y,
at
least,
because
none
of
the
mutations
in
the
Quintuple
stock
exhibits
sex
-linked
inheritance,
and
they
have
therefore
had
their
genes
assigned
to
an-
other
(the
second)
than
the
X
chromosome.
Duration
of
li
fe
in
this
respect
behaves
in
inheritance
exactly
like
the
morphological
mutations
with
which
it
is
associated
in
this
cross.
4.
The
F
t
fl
ies
are
significantly
less
variable
relatively
in
duration
of
life
than.
are
either
of
the
parent
stocks.
In
view
of
the
close
parallelism
between
inheritance
of
morphological
characters
and
inheritance
of
duration
of
life
which
has
so
far
been
noted
(and
which
will
be
shown
as
we
proceed
to
extend
to
many
other
genetic
relationships),
what
one
would
like
to
know
is
whether
the
F
t
's
are
less
variable
in
structural
characteristics
than
are
the
parent
forms.
We
have
no
data
on
this
point
for
the
fl
ies
with
which
we
have
worked.
F
2
GENERATIONS
Since
use
was
made
of
wing
form
(normal
wild
type
versus
vestigial
wing)
to
distinguish
the
fl
ies
morpho-
logically
in
this
generation,
attention
is
directed
fi
rst
to
the
Mendelian
segregation
of
this
structural
character.
Table
V
gives
the
data.
In
this
table
each
mating
is
TABLE
V
MENDELIAN
SEGREGATION
IN
F.
:
OF
NORMAL
AND
VESTIOUL
WING,
TABU-
LATED
BY
SEPARATE
MATINGS
Matins
Type
Numbers
of
F1
Flies
with
Wing
Normal
Vestigial
Probol,
I
lulu
Probable
Devi
'lion
of
-
of
Muting
Errors
R K
Such
.
fo.f‘tiou
K
ey.
is
1110
Trials
(5
)
PXP
55
03
7:1
Expected
47.25
75
15
56
3.8
1.04
(5)
P
X
P
146
202
75
Expected
151.5
$0
.5
±4.15
38.06
(3)
Total
202
63
265
73+75
Expected
198.7.5
66.25
265
*4.70
50
(3)
P
0
X.
R
115
10
125
76
Expected
93.'75
,
31.45
1.25
+3.27
.005
(3)
Z
0
,
X
P
9
I
08
34
132
77
Expected
99
33
152
±3.30
50
(3)
Total
213
1
44
257
70+77
Expected
192.75
60.25
257
*4.08
4.3
.37
(7)
Z
X
Z
115
10
134
72
100.5
Expected
3.9.5
13
4
*3.38
4.3
(7)
Z
X
Z
I
109
22
131
74
98
85
Expected
52.75
151
±3.34
3.2
3.00
(7)
Total
224
41
255
72+74
hxpected
198.75
66.25
265
*4.75
5.3
All
inter
s
Pi
matins
(
3
.
0
sad
7)
039
148
787
73+73+76
Expected
511025
196.75
787
+9.46
.074
+77+72+74
(2)
Vest.
X
Z
0
04
72
136
78
Expected
68
08
130
±3.03
50
(2)
Z
cf'
X
Vest.
9
I
24
44
68
70
Expected
I
34
34
68
±2.78
1.52
(2)
Total
1
88
116
204
78
+79
Expected
1
102
102
204
5.05
(6)
Vest.
cf"
X
P
C
'
73
70
82
Expected
72.5 72.5
145
*4.06
.0
50
No.
649]
STUDIES
ON
THE
DURATION
OF
TABLE
V
(Con.timwd)
Mating
Type
Number:,
c
f
F1
Flies
with
Winkj
!
I
Probable
Times
:
Probable
;
Deviation!
of
Occurrence
of
Mating
Normal
Vestigial
Errors
p
7E..
!Such
I
D
e
, *
a
Deviation
in
100
Trials
(6)
P
ob
X
Vest.
9
!
54
80
131
83
Expected
I
67
67
MB
*3.38
3.8
1.04
(()
Total
I
120 150
270
82
+83
Expected
139.5
199E
27.9
1.0
20.00
(2)
ali
back
and
croc8CS
of
and
1
(6)
Pi
on
Ve
lia
Y52
460
'
I.83
Expected
2
2
3
1
45
469
'
*7.30
10.55
(r)
wfla
el
X
Z
Y
I
142
0
142
80
Expected
.
144
0
142
1
--
(S)
z
0
,
x
wild
1
I
175
0
175
I
81
Expected
175
0
175
i
(2)
Wild
;,,
x
P
9
163
0
163
I
84
Expected
168
0
led
(2)
P
o
b
X
wild
9
262
0
202
85
Expected
.
902
0
202
TIN
NVOIJIPTIff
±VIT,T
16f
No.
649]
STUDIES
ON
THE
DURATION
OF
LIFE
163
given
separately.
The
probable
errors
of
the
Mende
-
halt
ratios
are
on
the
basis
of
Cum
as
the
standard
de-
viation
of
the
class.
The
actual
values
of
the
probable
errors
are
taken
from
the
extremely
useful
set
of
"Tables
of
Probable
Errors
of
Mendelian
Ratios"
is-
sued
by
the
Department
of
Plant
Breeding
of
Cornell
University,
In
general
the
observed
results
in
respect
of
the
in-
heritance
of
vestigial
wings
appear
at
first
glance
to
be
not
in
especially
bad
accord
with
Mendelian
expecta-
tion.
There
are,
however,
certain
systematic
diver-
gences
which
should
be
noted.
These
are:
1.
The
F,
inter
se
matings
(P
X
1
3
,P
X
Z
and
Z
X
Z),
when
taken
as
a
whole,
give
a
highly
improbable,
indeed
certainly
significant,
divergence
from
the
expected
3
:I
ratio.
Only
7
times
in
10,000
trials
with
samples
of
787
individuals
should
we
expect
in
random
sampling
to
get
so
great
a
deviation
of
fact
from
theory
as
that
actually
shown.
The
divergence
is
in
the
direction
of
too
few
vestigials.
This
divergence
is
due
more
to
P
X
Z
than
to
either
of
the
other
two
mating
types,
P
X
P
giving
indeed
a
significantly
non
-divergent
result.
2.
The
F,
back
crosses
upon
vestigial
give
a
more
probable,
but
still
by
no
means
satisfactory,
agreement
with
theory.
Only
about
once
in
10
trials
would
as
great
(or
greater)
deviation
as
that
realized
be
expected
in
random
sampling.
But
the
divergence
is
in
the
di-
rection
of
too
many
vestigials.
3.
'The
relations
stated
in
1
and
2
above
mean
that
whenever
the
genes
for
vestigial
were
contributed
to
F
2
zygotes
from
individuals
duplex
for
this
character
(but
in
the
back
crosses)
the
progeny
showed
an
excess
of
vestigial
as
compared
with
wild
type
fl
ies,
whereas
when
the
vestigial
genes
came
from
individuals
simplex
in
re-
spect
of
this
character
(as
in
the
1
7
,
inter
se
matings)
there
was
a
defect
of
vestigial
flies
in
the
progeny.
What
may
be
the
ultimate
meaning
or
explanation
of
this
fact
we
shall
not
discuss
now,
but
it
seems
advisable
to
point
it
out
as
a
fact.
TABLE
VI
Dr4421/
AND
SERVE
RESHIP
DISTRIBUTION
P03
W,u
TYPE
IF\
Fuzs,
DERIVED
FROM
Ster
so
MATTERS
07
r
j
,
P
X
P
(5)
P
XP
P
(5)
(
5
)
X
P
P
T
Z
xZ
9
XPR
(3)
(3)
PXZ
(3)
ZXZ
(
7
)
1
ZXZ
ZXZ
(
7
)
(7)
Mt
Pi
Inter
Se
Matiags
Mating Mating
TAM
Mating
Mating
Total
Age
in
Days
Mating
Mating
Total
72-77
73
75
76
77
72
74
(A
di
A'
(2
2
1
di di I/
1-
5
1
1
1,000
1
2
3
1,000
5
10
1,000
14
1,000
6-11
7 7
995
2 2
986
055
10
978
12-17
6
6
960
7
I
977
2
3
951
10
962
1
6
7
1
031
8
21
20
944
4
7
937
43
I
937
24-
99
1
I
13
14
806
7
31
808
15
18
06
63
870
30-35
5
3
8
827
17
13
30
062
5
826
46
772
36-41
2
13
15
787
13
!
13
26
521
12
15
1
790
56
700
42-47
17
18
35
713
16
13
29
300
23
61
.
723
125
012
48-53
15
24
39
540
8
10
18
I
263
17
34
'
451
91
416
9
29
38
347
22
9
31
I
178
1
10
21
299
90
274
50-65
3
20
23
158
3 3
33
1
9
1
22
205
al
133
4
4
45
0
I
-
1
1
5
1
6
21
107
26
53
72-77
3
2
5
25
0
0
0
0
I
3
13
8
13
78-83
o
o
0
0
-.
0 0
0
0
Totals
146
!
202
115
98
213
1
11
109
224
630
THE
AMERICAN
NATURALIST
No.
649]
STUDIES
ON
THE
DURATION
OF
LIFE
165
Turning
now
to
the
consideration
of
duration
of
life
in
the
F,
progeny
we
have
in
Tables
VI
and
VII
the
data
for
wild
type
F,
fl
ies
derived
from
inter
se
matings
(P
X
P,
P
X
Z,
Z
X
Z)
of
F1.
The
T
lines
for
the
three
total
groups
P
X
P,
P
X
Z,
and
Z
X
Z
and
the
original
wild
type
parent
are
shown
graphically
in
Fig.
2.
4000
\\-->
roc
\
co
QQ.
-
s
-
\
I0
1
0
6
/2
18
24
X
36
OZ
OB
56
60
66
TL
Y8
84
90
AGE
Al
DAYS
Pm.
Survivorship
lives
for
F2 WIld type
Elea
from
Fie
mated
inter
ae.
RIt
II
e=PxP;
dot
line
=
P
x
dusIddot
line
=
Z
x
Z;
solid
line
=
original
wild
type
parent
stock.
From
these
data
we
note
the
following
points
regard-
ing
the
wild
type
I:Vs
from
inter
se
matings
11
11
:
1.
Taking
all
such
fl
ies
together,
they
have
come
back
almost
exactly
in
duration
of
li
fe
to
the
same
mean
and
166
THE
AMERICAN
NATURALIST
[Vol,.
LVII
variability,
both
absolute
and
relative,
as
the
original
Old
Falmouth
parental
stock.
The
segregation
in
re-
spect
of
long
life
in
this
cross
appears
therefore
to
be
quantitatively
perfect,
and
to
have
followed
exactly
the
morphological
segregation
of
wing
form,
the
difference
of
the
constants
for
the
total
distribution
(last
line
of
Table
VII)
being
not
significantly
different
(or
indeed
approaching
significance)
from
those
of
the
Old
Fal-
mouth
parent
(first
line
of
Table
IV).
TABLE.
VII
BIOMETR1C
CONSTANTS
FOR
DURATION
OF
LIFE
OF
WILD
TYPE
F,
FLIES
num
inter
se
MATINGS
OF
IP,
Mating
Moan
Ago
YP
at
Death
Standard
:
Deviation
()loofas:lent
of
Variation
P
x
P
(Total)
(5)
46.94A
.73
days
9.201.01
days
83.00+1.21
I'
0
X
Z
9
(70)
.
(3)
:
39.16A
.83
days
3.263.59
days
33.80A1.67
Z
0
7
X
P
Q
(
77
)
(a)
35.89
±1.01
days
14.783.71
day:,
41.18±2.30
P
X
Z
(Total)
(3)
37.66A .65
days
1.1.083z9e
days
37.37+1.38
Z
X
Z
(Total)
(7)
40.01A
.73
days
10.19±.52
clays
35.10
±1,25
All
sneer
eunatingn.
(5)
-F(3)
43333
.42
day.
15.75
A.30
days
30.313
.77
H
-
(
7
)
2.
The
two
similar
-flier
se
mating
types,
P
x
P
and
Z
X
Z,
give
substantlilly
identical
results
for
duration
of
li
fe,
neither
means
nor
variability
differing
signifi-
cantly.
3.
The
cross
inter
se
mating
type,
P
x
Z,
has
a
sig-
nificantly
lower
mean
duration
of
li
fe
than
either
of
the
other
two
types.
The
differences
are:
For
P
x
P,
46.34-37.66
=
8.68
±
.98
days
For
ZX
Z,
46.01
—27.66
=
8.85
-I-
.98
days
In
variability
there
is
no
certainly
significant
differ-
ence.
The
corresponding
data
to
Tables
VI
and
VII,
but
for
vestigial
winged
F
2
's
are
presented
in
Tables
VIII
and
IX.
The
la,'
lines
for
the
three
total
groups
and
the
origi-
nal
vestigial
parent
stock
are
shown
graphically
in
Fig.
3.
TABLE
VIII
IDEATE
AND
SURVIVORSHIP
DISTRIBUTION
FOR
VESTIGIAL
F
2
FLIES,
DERIVED
FROM
inter
se
MATINGS
or
P
i
Age
in
Days
P
X
P
(
3
)
PcT
XZ
(
3
)
Z
P
XPQ
(
3
)
(3)
X
Z
Mating
ZXZ
(7)
1\18.
7
1,
4
ing
11I
FiThier
Matins
Mating
73
Total
Muting
70
M;
,1
ing
97
d
u
:
'rota]
Total
72-77
di
I'
d
u
'
di di
!
l
e
di di
11
duf
Ii
1-
2
3
4
1,000
2
1,000
2
7
1,000
13
1,000
3-
5
2
18
20
037
2
4
053
12
3
15
820
39
912
0-
8
3 3
610
864
1
463
5
041)
0-11
8 8
571
841
t
2
3
439
13
I
615
12-14
1
2
3
444
^50
2
366
7
514
1517
6
307
4
705
2
2
317
12
400
18-20
1 1
317
8
614
268
9
385
21-23
10
10
302
2
432
1 1
268
18
324
24-26
5
3
127
2
5
273
2 2
244
12
203
27-29
1
48
5
150
3 3
193
9
122
30-32
:32
1
45
3
3
122
4
61
33
35
1
:32
1
23
40
2
34
31
38
16
0
20
39-41
1
16
49
1
20
42-44
0
49
14
45
47
2
2
40
2
48-50
0
0
ToLti
7
56
63
10
34
44
19
22 41
148
17
210
ATOLL
1(2111C1
MILL
NO
SWRIBIS'
[669
168
SURVIVCES
THE
AMERICAN
NATURALIST
[You.
LITTI
1,000
\
\
100
\\
z..rz.
10
I -
0
6
12
Id
24
30
36
42
48
MI
AGE
IN
DAYS
rio.
3.
Survive
'ship
lines
for
Fe
♦estigial
flies
from
Iris
mated
inter
ee.
Dash
line
=P
x
P
dot
line
=
x
Z;
dashdot
line
=
Z
x
7.;
solid
line
=
original
vestigial
parent
st
dz.
From
these
data
we
note:
1..
Taking
all
vestigial
F.
flies
from
inter
se
matings
together,
the
distribution
of
mortality
agrees
in
an
ex-
traordinarily
close
manner
with
that
of
the
original
vestigial
parent
stock
(cf.
Tables
IIT
and
IV).
The
mean
duration
of
li
fe
of
the
original
vestigial
parent
stock
was
14.08
.23
days
and
that
of
the
F
2
vestigials
from
inter
se
matings
is
14.60
4-
.57
days.
The
segrega-
Na.
649]
STUDIES
ON
THE
DURATION
OF
LIFE
169
ton
in
respect
of
mean
longevity
is
plainly
exact
and
precise.
The
variability,
both
absolute
and
relative,
is
probably
significantly
higher
in
the
F,
than
in
the
P
fl
ies
(vestigials),
but
the
F,
samples
are
not
large
and
too
much
stress
is
not
to
be
laid
upon
these
differences.
TABLE
IX
1310D111PRIC
CONSTANTs
ii•Ox
DirenTiorc
OF
LiFE
oP
VESTTGW.
F,
FLiEs
FROM
imer
so
MATINGS
OF
F,
Mating
.
cype
Mean
atD
Age
eat
'
S1
-
iidml
Deviation
P
X
P
(Total)
(L)
13.111±
.70
days
9.25±
.54
days
P
e x 1
9
(7
6).
..
(3)
.
17.45+2.18
days:
10.23+1.54
days
Z
0
7
X
P
9
(77)..
.
(3)
!
18.81
.8.9
days
!
7.
093
.03
clays
P
X
Z
(Total)
(3)
i
18.50±
.85
days
8.34A
.60
days
(Total)
(7)
!
13.15+1.31
days'
12.40A
.93
days
All
inter
se
matings.
(5)
+(3)
i
11.00
.57
days
10.31+
.40
clays
+(7)
!
!
!
I
Coctlicieut.
of
Variation
70.70
±
6.01
58.01
+11:19
4098±
3.85
45.00+
3.85
91.78+11.81
70.05
3.91
2.
The
simi
ar
mating
types
(P
X
P
and
Z
X
Z)
give
identical
resu
is
in
mean
and
variability
(having
re-
gard
to
the
probable
error
involved)
in
respect
of
dura-
tion
of
life.
This
agrees
with
the
previous
findings
for
the
case
of
wild
type
F,
fl
ies.
3.
The
inter
se
cross
mating,
1'
x
Z,
gives
vestigial
progeny
fl
ies
having
a
significantly
higher
mean
dura-
tion
of
life
than
either
of
the
other
two
types
(P
X
P,
Z
X
Z).
The
differences
with
their
probable
errors
are:
For
P
X
P,
18.50-13.10
=
5.40
±
1.16
days
For
Z
x
Z,
1R.50
13.15
=
5.35
±
1.56
days
While
these
are
not
as
large
as
the
reverse
differ-
ences
found
above
in
the
case
of
the
wild
type
F.
fl
ies,
nevertheless
they
may
be
safely
regarded,
we
think,
as
statistically
significant.
The
bearing
of
these
differences
between
the
several
mating
types
upon
the
general
problem
of
the
inherit-
ance
of
duration
of
life
will
be
discussed
further
on.
The
data
regarding
duration
of
li
fe
in
the
F,
individ-
uals
produced
by
back
crossing
F
1
upon
the
parental
types
are
given
in
Tables
X
and
XI
for
wild
type
fl
ies,
and
Tables
XIT
and
XIII
for
vestigial
fl
ies.
TABLE
X
DENTE
AN
SURVIVOREDIP
DISTRIBUTIONS
FOR
la'
ILO
TYPE
F
2
FLIES,
DLINNED
FROM
BACK
CROSSES
OF
F.
ON
THE
PARENTAL
STOCKS
Age
Days
Z
X
Old
(8)
(Ree
Falmouth
proeallY)
P
X
Old
(4)
(Reciprocally)
Falmouth
Total
All
Back-
oros4es
F1
on
Wild
Type
Mailings
80+81+84
(2)
Ma-
hbg
I
Z
X
Vestigial
(Recipmea(ly)
Ma-
ling
Total
P
X
Vestigial
(6)
(Reciprocally)
All
Baek-
ernsses
of
on
yestigim
Ma-
tiag.
Ma-
tins
Total
Ma
-
1
Ma-
ting
1
tiag
Ma-
Ma-'
ting
.
ting
Total
Mati44.54
78
-870
+82
80
81
84
+85
78
79
61
i
83
+83
1
di
di di
I,
!
di
1
11
di
I
d.'
'
di
d
i
'
11
1-5
6
7
13
1
1,00(1
2
,
1,1100
15
1,000
3 3
I
1,000
1
-
1,000
1.000
6-11
5
28
33
959
1
.
25
26
005
fns
966
3
8
1
1,000
10
086
12-17
18-23
34
11
55
565
849
4
13
5
11
9
24
923
890
7
1:1
8
89
74
1
5
5
3
3
8
943
3
0
852
2
13
8
1
10
938
7
!
20
800
18
26
040
857
21-211
10
24
40
075
51
20
77
833
117
760
10
3
13
784
18
6
1
24
705
37
737
30-35
35
19
51
540
13
13
36
622
90
588
15
2
17
636
17
18
35
519
52
567
36-41
7
13
20
379
20
15
35
523
55
1
156
10
4
20
443
3
0
0
248
29
327
42-47
10
12
22
315
10
28
44
427
60
I
37.5
7
2
9
216
4
4
8
178
17
104
4S-53
15
13
28
146
11
18
29
307
57
279
4
1
5
114
3
3
110
8
115
54-59
3
28
31
158
10
63
00
227
100
1
195
3
1
4
7
1
8
93
12
78
00
-05
,
9 9
18
60
7
4
11
38
29
1
48
1
11
2
1
3
31
4
23
60
71
1
1
3
1
2
3
8
4.
6
0
1
1
8
1
5
72-77
.
_
0
0
0
Totals.
142
175
(117
163
202
365
632
04
24
88
75
54
129
217
THE
AMERICA
N
NATURALIST
No.
6191
STUDIES
ON
TILE DURATION OF
LIFE
171
TABLE
XI
BloatETRIC
CONSTANTS
FOR
1
IfT0.T.LON
OF
TAPE
OF
WILD
'T
YPE
FLIFS
PROM
Bacx
CROSSES
Malings
Typo
Men,.
Age
at
Death
Sta.Klaid
(Medi
rient
Oi
Deviation
Variation
Wild
6
1
X
Z
Q
(80)
(8)
32.50+
.83
days
14.07+
.52
days
45.07+2.14
Z
r7
X
wild
9
(81)
(8)
33.01+
.03
clays
18
8
.19.E
.60
days
53.64+2.43
Total
Z
X
w➢d
(8)
33.30+
.63
days
16.72+
.45
days
5120+1.65
Wild
di
X1'
9
(84):
(4)
30.57+
.68
days
12.85+
.48
days
35.13+1.47
P
X
wild
9
(85)
(4)
38.83+
.82
days
17.31
+
.58
days,
44.50W1.77
Total
X
wild
!
(4)
37.82+
.55
clays
15.52+
.30
days
41.03+1.18
Total
back
cross
on
wild
(8)
-
K
4
)
85.72+
.42
days
10.25+
.30
days!
45.48+
.99
Vestigial
sd
X
Z
9
(2)
85.0
+
.111
days
10.81+
.04
days"'
30.34+1:07
Z
od
X
Vestigial
9
(79)
(2)
26.56+2.11
days
15.20+1.49
days'
57.57+7.23
Total
Z
X
Vestigial
(2)
33.15+
.02
days
12.85+
.05
days
38.75+2.25
Vestigial
cf
X
P
4
(82)
i
(0)
82.52+1.00
days
14.05+
.77
days
43
22+2.79
P
6'
X
Vestigial
9
(83)
(6)
29,33*1,05
days,
11.40+
.74
days
30.05+2.00
Total
P
X
Vestigial
(6)
31.10+
.78
days
13.12+
.55
days
43.08+2.06
Total
back
cross
of
vestigial
(
2
)
-1(6)
31.08+
.60
days
13.05+
.42
days
40.80+1.5a
From
Tables
we
note
the
following
points:
1.
There
is
evident
a
general
degradation
of
mean
duration
of
life
ill
back
-cross
fl
ies,
whether
compared
with
original
parents,
P
i
's
or
inter
se
F
e
's.
Thus
we
have
the
following
system
of
differences:
Original
wild
pan
ont.
-
all
wild
type
flies
from
bark
era
saes
of
F,
on
Wild
44.26
-
35.72
=
8454
-1
.61
Original
wild
parent
-
all
wild
typo
!fl
ies
from
back
crosses
of
F,
on
vestig-
ial
=44.26
-31.98
-
12.28
±
.71
Original
vestigial
parent
all
vestigial
flies
from
hack
grosses
of
F,
on
vostigial
=
14.08
-13.84 =1.24
+
.36
There
can
be
no
doubt
about
a
significant
lowering
of
mean
duration
of
life
of
these
back
-cross
fl
ies
as
com-
pared
with
their
original
progenitors
in
the
P
genera-
tions.
In
the
worst
case
(the
last)
the
difference
is
3.4
times
its
probable
error.
In
the
case
of
F,
fl
ies
it
is
idle
to
take
the
differences
here
because
the
F,'s
are
longer
lived
(as
we
have
al-
ready
seen)
than
the
wild
type
original
parent.
There-
fore
the
differences
between
F
1
and
back
crosses
must
of
necessity
be
larger
than
those
we
have
just
examined.
TABLE
XTI
DEATH
AND
SURVIVORSHIP
DISTRIBUTIONS
POE
VESTIOP,T,
F,
FURS,
DRAFTED
FROM
BACK
CROSSES
OF
P,
ON
QUIFIFIRLF,
STOCK
Age
in
Days
1-
2
3-
5
6-
8
9-11
12-14
15-17
18-20
21-23
24-26
9
7-2
30-
4
1
33-35
36-38
39-41
7
,
X
Vestigial
P
X
Vestigial
All
Bark
Tosses
of
(2)
(Reciprocally)
(0)
(Reciprocally)
Fi
on
Vestigial
Mating Mating
Mating
I
Mating
78
79
Total
82
83
Total
r2,'
di
1
1,
4
1
Totals
72
2
12
6
1
3
1,000
1
2
18
974
5
14
I
9
819
10
22
18
741
15
16
17
586
19
15
18
440
8
4
7
284
8
4
10
224
2
6
138
1
86
1
1
26
1
17
--
1
1
1
--
I
0
44
116
70
80
150
Matings
78+70
+82+83
dd
Ti
All
Pi
rP
:
All
Vestigial
X
Ve
-
L.
9
:
c
7
,
-
X
Fi
9
Main'.
Ma
ings
79-c83
,
78+82
di ii
.
di
3
1,000
19
980
32
,
853
:31
010
34
!
433
12
207
12
4
20
13
0
6
37
41
49
51
30
19
14
2
1
206
1,000
977
838
684
500
308
195
124
71
46
15
11
4
0
4
26
28
23
21
4
1
1
1
1,000
I
2
068
11
756
13
532
26
317
30
177
23
121
!
12
65
10
32
6
24
7
16
1
16
1
8
01
-
1,000
986
908
817
634
423
261
176
106
63
14
7
0
124
142
LSITra
ILITAT
_NY
OVIFIlliti
P1111
ci
No.
649]
STUDIES
ON
TEE
DURATION
Oh'
LIFE
173
TABLE
XIII
BlOmnsc
CONSTANTS
FOR
TEGRATTON
OF
LIFE
OF
VESTIGIAL
P,
FLIES
FROM
BACK
CROSSES
Mating
Mean
Age
Standard
TYpe
at
Death
Deviation
Coefficient
of
Variation
Vestigial
o
9
X
Z
9
(78)
(2)
10.47,4.59
clays
7.
41±.42
days
45.01
n3.00
Z
0
X
Vestigial
9
(79)
(2)
11.41
n.78
days
7.61
n.55
days
06.0230.63
Total
Z
X
Vestigial
(2)
14.55
n.49
days
7.893A5
days
54-223.3.03
Vestigial
a'
X
P
9
(82)
(6)
12.80
n.12
days
5.211n.311
clays
40.4433.66
P
d
X
Vestigial
R
(83)
(0)
10.23
n.11
days
52.06+3.45
Total
P
X
Vestigial
(0)
11.51±.30
days
5.44
L.21
days
47.30E2.22
Total
back
crosses
on
Vestigial
(2)
-9(6)
12E1
n.28
days
0903.20
days
52.91±1.93
Total
Fi
co
X
Ves
i
t
. .
10.71
n.38
clays
0.29
n.27
days
58.75n3.27
Total
Vestigial
o
f
X
Fi
9
14.00
n.38
days
0.07±27
days
15.38±2.16
.•-
.
For
the
F
2
's
from
inter
se
matings
of
F
1
we
have
the
following:
F,
(from
P
x
P)
wild
type
-wild
type
from
back
cross
P
X
wild
=46.31
-
37.82
=
8.52+
.91
F,
(from
Z
X
7)
wild
typo
--
wild
type
from
back
GIES,
Z
X
wild
=
46.01
-
33.30
=12.71
-4-
.96
F,
(from
P
x
7)
wild
type
-
wild
typo
from
all
bark
,'rossoa
on
wild
=.37.66-35.72=
1
94
-d
.77
F.
(from
P
x
p)
wild
type
-wild
type
-
Flom
hack
cross
P
X
vestigial
=
46.34
-
31.19
-
15.15
-9
1.07
P
a
(from
P
X
Z)
wild
type
--
wild
type
from
all
back
crosses
on
vestigial
=46.01
-
33.15
=12.86
-1-
1.17
IL
(from
P
X
7)
wild
type
-
wild
type
from
all
back
crosses
on
vestigial
=
37.68
-
31.98
=
5
68
.8-
.88
Thus
it
is
clear
that
the
wild
type
F
2
fl
ies
from
back
crosses
of
F,
upon
either
original
parent
have
a
signifi-
cantly
lower
mean
duration
of
life
than
do
the
F
2
'fl
ies
from
inter
se
matings
of
F,.
The
only
exception
to
this
rule
is
found
in
the
ease
of
the
F.,
from
the
P
X
Z
mat-
ings,
when
compared
with
all
back
crosses
upon
wild.
This
difference
is
less
than
three
times
its
probable
error.
Turning
now
to
vestigials
we
have:
174
THE
AMERICAN
NATURALIST
[You.
LVII
P,
(from
P
x
P)
yes
iglal
-
vestigial
from
back
cross
P
vestigial
-13.16 -11r1 1.59*
.54
P,
(front
Z
x
Z)
vestigial
-
vestigial
from
back
cross
Z
X
vestigial
=13.15
-
14.55=
-
1.40
x
1.40
F.
(from
P
X
Z)
vestigial
-
totstigial
from
all
back
crosses
ou
vostigiol
=18.50
-12.94
=
5.66
x
.89
Here
the
only
difference
which
is
significant
is
the
last,
which
invokes
the
P X
Z
P
t
fl
ies.
Everywhere
this
group
gives
an
aberrant
result.
2.
There
are
certain
definite
and
orderly
differences
in
mean
duration
of
life
between
the
fl
ies
from
the
two
moieties
of
otherwise
identical
reciprocal
back
crosses,
and
&so
between
the
flies
coming
from
back
crosses
in-
volving
the
two
sorts
of
F,'s,
namely
P
and
Z.
These
differences
are
as
follows:
WILD
PYIW
d
d
ieWS
(wild
x
7
7)
-
(Wild
d
X
P
=32.56
-36.57
4.01
1.07
(Z
5
/
Wild
9)
-
(Pd
X
Wild
2)
=33.91
--
38.83
=
--
4.92
X
L24
(Vest.
5
/
z
2)
-
(Vest.
5 x
P
2)
=35.63
-32.52
=
3.11
-X
1.42
(z
5
x
vest.
9)
-
(P
d
X
vest-
9)
=
26.56
-29.33 =-
2.77
±2.36
VESTIGIAL
FLIES
(Vast.
cf
X
Z
-
-
(Vest.
A
X
P
=16.47
-12.86
_
3.61
X
.72
(Z
X
vest.
2)
-
e
X
vest.
2)
=11.41
-10.32
=
-1
1.08
X
.88
From
these
differences
it
appears that
when
the
back
-
cross
mating
is
such
as
to
produce
only
wild
type
flies
the
offspring
are
significantly
longer
li
ved
when
the
F,
parent
is
of
P
origin
than
when
it
is
of
Z
origin.
The
reverse
appears
to
be
true
when
the
mating
is
of
the
sort
to
produce
both
wild
and.
vestigial
types
in
the
off-
spring,
though
most
of
the
series
here
are
unfortunately
too
short
to
give
statistically
significant
differences.
The
differences
between
reciprocal
matings
are
as
follows:
WILD
MISS
FLIES
(Wild
2X
ZP)
-
(Z5
X
wild
2)
=32.56
-33.91
= -
1.35
ci
1.25
(Wild
e
X
r -
(Pcr
X
wild
2)
=36.57
-38.83
= -
2.26
X
1.07
(Vest.
di
x
Z
(Z
g
x
vest.
9)
=35.66-26.56
= +
9.07
si
2.30
(Vest.
?)
-
(1
)
x
vest.
2)
=32.92
-
29.33
= +
3.19
*1.5i
VESTIGIAI.
FIATS
(Vest5XZP)-(2.dX
vest.
2)
m:16.47
-11.41
=
-1
5.96
-it
.98
(VestdXP2)
-
(P
d
i
X
vest-
9)
=
12.86
10.33
= +
2.53
-s
.49
No.
649J
STUDIES
ON
THE
DURATION
OF
LIFE
175
From
these
data
it
appears
that
in
back
crosses
of
on
wild
the
offspring
tend
to
be
longer
lived
(though
because
of
the
shortness
of
the
series
neither
difference
is
significant)
when
the
mother
is
pure
wild:
in
back
crosses
of
F,
on
vestigial,
the
offspring,
whether
wild
or
vestigial,
are
shorter
lived
when
the
mother
is
pure
vestigial.
All
the
differences
but
one
are
clearly
sig-
nificant,
and
that
one
is
probably
so.
There
thus
ap-
pears
here
a
definite
matroclinal
tendency
relative
to
the
inheritance
of
duration
of
life.
F
s
GENERATION
The
matings
made
to
produce
1',
flies
were
few
in
number,
as is
shown
in
Table
It
and
covered
only
a
small
fraction
of
the
possibilities.
In
Table
XIV
are
given
the
Mendelian
results,
relative
to
the
segregation
of
normal
and
vestigial
wing,
for
all
the
matings
it
was
possible
to
use.
The
other
matings
indicated
in
Table
II
could
not
be
used
in
the
analysis,
because
they
were
mass
matings
in
which
a
fly
might
be
either
heterozy-
gous
or
pure
dominant,
and
as
a
matter
of
fact
some
were
heterozygous,
as
indicated
by
the
results,
in
the
case
of
the
three
matings
not
included
in
Table
XIV.
Because
of
interest
in
their
apparently
aberrant
be-
havior,
already
brought
out,
attention
was
confined
in
making
up
the
matings
for
F,
to
the
IF,
produce
of
P
X
Z
matings.
It
will
be
seen
from
Table
XIV
that
in
every
case
of
matings
involving
vestigial
F
2
females
(which
includes
all
matings
in
the
table
except
the
last)
there
is
an
ex-
cess
of
vestigial
flies
in
the
offspring;
over
what
would
be
expected
on
Mendelian
theory.
The
discrepancies
are
certainly
significant
in
the
totals,
and
also
for
two
out
of
the
four
separate
matings.
It
will
be
recalled
that
Table
V
shows
an
excess
of
vestigials
from
the
back
crosses
of
F,
on
the
parent
stocks.
Tables
XV—XVIII
inclusive
give
the
F,
results
as
to
duration
of
life.
176
THE AMERICAN NATURALIST
[VOL.
iLVII
TABLE
XIV
MENDELIAN
SEGREGATION
IN
F
a
OP
NORMAL
AND
VESTIGIAL
WINGS,
TABU-
LATED
BY
SEPARATE
MATINGS
Ma
Vestigial
9 9
Number
Flies
with
Nor-
mad
of
Fs
Wi
sgs
Ves-
To
-
tigial
tal
Prob-
able
1
able
15
"
1!
Deviati„„
Probable
Times
of
(lent
of
Such
a
Deviation
in
100
Trials
Ma
-
tins
No
'
P
.
P
.
I
De
8
.!
j
(54)
Z
33
X
1)
(70)
9
76
121
197
SO
Expected
98.5
95.5
107
±4.70
4.8
.12
(54)
Z
3
7
X
D
(77)
4
67
76
143
go
Expected
71.6
71.5
143
8
-4.03
1
1.1
45.81
(54)
Total
143
107
340
0
890
IC40)4014(1
170 170
340
6.'.2
9
I
4.3
s'
37
(52)
P
e
X
D
(76)
2
05
107
202
'
SS
Expected
101
101
002
*4.79
1.3
(5
2
)
P
3
7
X
1)
(77)
2.
63
110
182
92
Expected
...
.
.
01
41
182
24.55
6.2
.005
(52)
Total
158
220
384
8-802
Expected
199
193
1384
6.3.1
5.1
.07
Wild
type
9
0
(52)
P
e
x
D
(76)
9
:
17
5
0
1175
89
Expected
175
0
175
-
.
TALLE
XV
DEATH
AND
SURVIVORSHIP
DISTRIBUTIONS
FOR
ILD
TYPE
Fp
FLIES
Z
Fl
0
7
X
Vestigial
D
Fs
9
P
Fl
1P
X
Vestigial
D
Fs
9
I'
E
l
o'
X
Wild
Type
DF
2
Age
Ma-
Ma.-
Ma-
1
Ma-
ting
tine
'I'otal
ting
Sing
To
2x1
Mating
Days
80
00
HE
1
92
89
di di di
14
di
/,'
di
1
I
s
'
1
1
2
1,000
11
I
3
14
1,000
2
1
1,000
1
7
086
6
8
14
011
6
:
051)
12-17
0
13
22
037
0
:
6
10
823
26
!
954
18-23
.
18
24
12
783
21
9
30
722
35
806
24-20
5
11
10
41)0
13
I
4
17
532
12
606
30-35
36-41
5
6
0
14
378
S
280
6
r
(
5
1.
I
6
11
12
421
354
23
14
537
406
42-47.
0
1
7
224
6
4
I
0
278
29
:120
48-53
2
17
I
175
14
12
215
18
100
54-59
0
2
I
8
56
3
5
51
10
57
60-65.
.
0
0 0
Totals
1
76
67
143
95
1
63
158
176
No.
649]
STUDIES
ON
TILE
_DURATION
OF LIFE 177
TABLE
XVI
Blompasalc
CONSTANTS
FOR.
DURATION
Or
TAFE
op
Wan
I
OPE
1S
3
FLIES
Mating
Moan
Age
I
21e
.
at
Death
I
Standard
Deviation
Coallicion1
of
Variation
Z
Fr
:I
X
Vest.
Dm
F,
(80)•
(54)
32.53±1.22
15.741
.86
48.3713.21
111
H
X
Vest.
1/2-
F,
9.
(90).
(51)
21.68
.87
10.55±
.61
42.8012.02
Total
Z
F,
P
X
Vest.
1)
P
(54)
28.851
.80
11.11
±
.50
48.91±2..37
P1
1
,
an
X
Vest.
Dm
F,
9
(8S)
(52)
27.2511.10
15.8/41
.78
,
5/4.29
3.70
P
ID
P
X
Vest.
Ds
F,
9
(92).
(52)
31.40±1.44
16.00
11.02
I
53.8314.07
Total
P
an
X
Vest.
1)
9
(52)
28.00±
.88
16.421
462
,
56.8212.77
P
R e
X
Wild
type
11
22
Fe
e
(SO)
32.05±
.72
14.181
.51
44.3611.88
TABLE
XVII
DEATH
AND
SURVIVORSHIP
DISTRDSUTTONS
FOR
VESTIGIAL
F,
FLIES
Age
in
Days
Z
F,
di
X
Vestigial
D
F2
P
FI
Mating
88
.1
X
Vestigial
Mating
02
D
F,
9
To.
al
Mating
1
:
Slating.
86
.
00
Total
,
a
El.'
4,'
d2
212
12
1-
2
46
12
1,000
22
21.
43
I
1,000
3-
5
so
25
7
16
27
810
6
7
,
21
330
12
10
31
619
9-11
9
I
7
11
188
14
29
482
12-14
5
I
4
107
22
34
56
I
354
15-17
1
7
01.
16
3
19
I
100
18-20
20
2 2
22
21-2
4
-
I
-
11)
13
21
26
1 -
1
10
13
27-29
-
--•
13
30
32
0
13
33-35
13
36-38
9
Totals
121
70
107 107
0
220
TABLE
XVIII
1310KSWilc
CONSIANTS
POE
DVNATION
Of
LINE
Or
VEsTSNAL
F,
FISIES
Mating
'Type
Mean
Mean
Age
Death
Standard
Coefficient
of
Deviation
Variation
Z
E,
27'
X
Vest.
Dm
F,
9
(86).
(54)
'
5.28±.28
4.04
-m.20
87.95±0.09
Z
10.
P
X
Vest.
DTI
F,
0
(00).
(54)
7.0732.33
4.211.23
59.6014.26
Total
Z
Fi
X
Vest.
1)
9
(54)
5.97
4.57
76.40±3.83
P
Fr
(7.
X
Vest.
Dm
9
(88).
(52)
:
0.03,6.45
6.98
70.26±4.57
Plr,
os
X
Vest.
Dm
F2
9
(02).
(52)
8.17±18
55.29=1-3.07
Total
P
1i'
1
X
Vest.
1)
9
(52)
9.001.26
5.814±.19
65.3012.82
178
THE
AMER/CAN
NATURALIST
Mat
LVII
The
first
point
which
will
be
noted
from
these
Tables
XV—XVIII
i.s
the
still
further
marked
degradation
in
duration
of
life
of
F,
flies,
as
compared
with
those
of
earlier
generations.
None
of
the
means
approach
in
value
the
original
parental
stocks,
and
are
even
well
be
-
/ow,
taken
as
a
whole,
the
hm,
fl
ies
from
back
crosses.
Another
result
is
the
system
of
differences,
in
the
main
significant,
in
duration
of
life
according
to
whether
the
F,
vestigial
females
used
were
from
mating
76
or
mating
77.
D„
females
came
from
Zdx
P
and
D77
females
came
from
PSXZ
2.
Now,
when
a
ll
,„
9
,
was
mated
back
to
Z
cr,
the
resulting,
wild
type
fl
ies
were
longer
lived
and
the
vestigials
were
shorter
lived
than
when
a
D„
female
was
mated
to
a
Z
6
1
.
When
these
same
two
sorts
of
females
were
mated
to
a
P
g
the
off-
spring
of
the
Dn
females
were
longer
lived
if
normal
winged
and
shorter
lived
if
vestigial
winged
than
the
offspring
of
the
B
70
females.
In
the
one
mating
(89)
which
enables
a
comparison
of
D
70
wild
type
and
a
ti
n
,
vestigial
each
mated
to
the
same
type
of
male
(F),
the
resulting
wild
type
progeny
are
longer
lived
from
the
wild
type
mother.
The
differ-
ence
4.80
±
1.31.
days
is
almost
certainly
significant.
THE
EFFECT
OF
ABSENCE
OF
WINOS
ON
DURATION
OF
LIFE
It
can
not
fail
to
occur
to
one
that
perhaps
the
reason
why
vestigial
winged
fl
ies
are
invariably
short
li
ved
is
merely
and
directly
physiological,
arising
because
such
a
fly
has
no
junctional
wings
and
therefore
can
not
fl
y,
but
has
its
activity
confined
to
walking
and
jumping.
Stated
in
extreme
forms,
the
idea
would
be
that
so
far
as
inherent
constitutional
factors
influencing
duration
of
life
are
concerned
Quintuple
and
wild
stocks
are
on
the
same
footing,
but
because
Quintuple
has
no
func-
tional
wings
it
is
physiologically
unable
so
to
conduct
its
life
as
to
realize
its
constitutional
potentialities
in
respect
of
longevity.
It
is
possible
experimentally
to
test
this
idea,
in
part
No.
649]
STUDIES
ON
THE
DURATION
OF
LIFE
179
at
least.
Suppose
in
a
large
number
of
wild
type
flies
immediately
after
emergence
the
wings
are
cat
off
close
to
the
body.
Since
they
do
not
regenerate
these
fl
ies
must
then
go
through
life
wingless.
If
then
a
duration
of
life
experiment
is
carried
through
with
them,
and
a
life
curve
constructed,
will
this
curve
be
like
that
for
Quintuple
fl
ies,
or
will
it
be
like
that
for
normal
wild
fl
ies?
Or,
put
in
another
way, what
difference
in
the
distribution
of
mortality
of
Drosophila
does
the
pres-
ence
or
absence
of
wings
make
in
a
purely
physiological,
somatic
sense?
This
experiment
we
have
carried
out
on
a
large
scale
twice.
Unfortunately
we
can
give
here
the
result
only
of
the
second
of
these
investigations,
because
in
the
case
of
the
first
series
all
the
surviving
fl
ies
were
destroyed
when
rather
more
than
half
way
through
the
life
curve,
owing
to
the
accidental
misbehavior
of
the
regulating
apparatus
of
the
incubator
in
which
they
were.
The
results,
however,
so
far
as
they
went
wore
in
entire
agreement
with
those
here
presented.
The
fl
ies
in
this
experiment
emerged
April
16-19,
1922,
from
25
mass
matings
(2
pairs
each)
Old
Fal-
mouth
stock,
line
107,
generation
26.
Progeny
were
taken
from
the
mating
bottles
at
24
-hour
intervals
and
etherized.
Only
the
fl
ies
whose
wings
were
fully
ex-
panded
and
dried
were
used,
in
either
experimental
or
control
group.
With
fi
ne
scissors
both
wings
were
clipped
off
as
close
to
the
thorax
as
possible,
care
being
taken
to
avoid
injury
to
the
halteres.
The
experimental
and
control
bottles
--with
30
fl
ies
in
each
bottle
—were
made
up
alternately
to
distribute
the
fl
ies
at
random.
Both
sots
were
then
put
through
a.
standard
duration
of
life
test
at
25°
C.
in
the
usual
way
in
which
all
our
work
is
done,
with
the
results
shown
in
Tables
XTX
and
XX
and
Fig.
4.
From
the
results
it
is
evident
that
surgical
removal
of
the
normal
wings
of
wild
type
fl
ies
at
the
time
of
emer-
gence
somewhat
shortens
duration
of
life,
but
to
no
such
180
THE
AMERICAN
NATURALIST
[VOL.
LVII
TABLE
XIX
DEAITI
AND
SURVIVORSHIP
DISTRIBUTIONS
von
WILD
TYPE
FLIES
(a)
nom
WINGS
SURGICALLY
REMOVED,
AND
(b)
WITH
NORAJAL
WINGS
Age
in
Days
Normal
Wings
()A
1-
3
5
I
1,000
4
1,000
4-
6
2
000
10
902
7-
0
4
986
20
973
10-12
4
970
9
934
13-15
9
971
34
917
12
053
30
852
19-21
22
930
45
704
22-24
22
887
38
707
25-27
17
844
26
634
28-30
33
811
37
584
31-33
21
747
21
513
34-36
20
706
22
472
37-30
20
007
19
430
40-42
33
1,18
21
303
43-45
34
553
20
353
46-48
34
480
24
393
49-51
33
420
20
256
52-54
42
350
32
218
56--57
64
274
28
150
58-60
25
150
27
102
20
101
1
50
64-66
12
62
29
67-6')
39
12
70-72
8
29
10
73-75
2
14
4
76-78
4
0
79-81
0
82-84
85-87
0
0
'Totals
514
519
TABLE
XX
BIOAIETRIC
CONSTANTS
FOR
DURATION
OF
LIFE,
IN
WILD
TYPE
FLIES
(a)
WITH
WINGS
SURGICALLY
REMOVED,
AND
(b)
WITH
NORVAL
WINOS
Groan
Mean
Duration
of
Life
Standard
Deviation
Coefficient
of
Variation
Wings
removed
4500.50
days
17.01+.36
days
49.23-1.20
Normal
wings
43.26±A7
days
15.0042.33
days
30.27±
.36
extent
as
would
he
required
to
account
for
the
normal
difference
between
wild
and
Quintuple
stocks
in
this
re-
spect.
The
differences
between
clipped
and
unclipped
groups
are
significant
in
comparison
with
their
prob-
No.
6401
STUDIES
ON
THE
DURATION
OF
LIFE
181
able
errors,
but
the
difference
between
the
means
is
only
about
one
fourth
of
the
difference
between
wild
and
Quintuple
means.
e
cnd.
Ni~
rd.
cs
nn
..
<
3
<88..
8N
/00
.5UEVIV<WS
1
I I I
0
6
12
18
24
20
36
42
d8
6.4.
60
66
72
73
84
.90
AGE
IN
DAYS
FIG.
4.
Survivorship
Hues
for
normal
wild
type
flies
(solid
lines)
and
flies
of
same
stock
with
whego
removed
at
emergence
(broken
lines).
It
may
be
safely
concluded
that
the
difference
between
wild
and
Quintuple
stocks
in
respect
of
duration
of
li
fe
is
not
solely
due
to
the
fact
that
the
Quintuple have
no
functional
wings
while
the
wild
do.
One
perhaps
could
reasonably
have
concluded
this
on
general
grounds,
but
it
seemed
well
to
have
an
experimental
proof.
Cer-
tainly
a
considerable
part
of
the
difference
between
182
THE
AMERICAN
NATURALIST
[Von..
LIT11
clipped
and
unclipped
shown
in
the
data,
and
perhaps
nearly
all
of
it,
is
due
to
the
unavoidable
and
unper-
ceived
gross
injury
to
the
.
fly
at
the
time
of
the
opera-
tion.
Three
times
as
many
clipped
as
unclipped
died
during
the
fi
rst
10
days
of
life.
No
effort
was
made
at
the
operation
to
do
anything
to
the
wing
stumps
to
pre-
vent
loss
of
body
fl
uids.
We
have
in
the
results
the
maximum
effect
of
both
absence
of
wings
throughout
life,
and
a
rather
rough
major
surgical
operation
at
the
start
of
life.
If
the
second
of
these
two
factors
could
he
eliminated
and
leave
the
fi
rst
by
itself
alone,
it
is
probable
that
the
two
curves
of
Fig.
4
would
be
insig-
nificantly
different
from
each
other.
Discussion
of
Results
So
far
we
have
studiously
avoided
any
discussion
of
the
meaning
of
the
results,
desiring
to
present
the
ob-
jective
facts
separated
clearly
from
any
interpretation
of
them.
We
wish
now
to
enter
-
upon
a
brief
discussion
of
the
meaning
of
what
we
have
found,
prefacing
our
remarks
in
this
direction
by
the
statement
that
we
re-
gard
this
investigation
as
in
no
way
fi
nal,
but
rather
as
preliminary
to
a
series
(probably
long)
of
experimental
studies
which
must
be
made
before
anything
like
a
com-
plete
understanding
of
the
mechanism
of
inheritance
of
duration
of
life
in
Drosophila
will
be
possible.
For
convenience
in
discussion
we
have
summarized
all
our
experiments
in
the
form
of
pedigree
diagrams
hi
Figs.
5-17,
entering
beside
each
designation
of
the
type
of
fl
y
the
mean
duration
of
life
of
that
group
to
the
near-
est
whole
number
of
days.
With
the
data
before
us
we
may
point
out
the
broad
results
of
the
investigation.
These
seem
to
us
to
be:
1.
Duration
of
life
behaves
in
the
F
i
,
F
2
and
a
part
of
the
possible
P,
offspring
from
reciprocal
crosses
of
short-lived
and
long-lived
strains
of
Drosophila
as
any
character
of
the
organism
would
be
expected
to
behave
if
it
were
inherited
in
a
Mendelian
manner.
There
is
a
No.
049]
STUDIES
ON
THE
DURATION
OP
LIFE
183
somewhat
higher
duration
of
life
in
F,
than
in
either
of
the
parent
stocks
(effect
of
fi
eterosis).
There
is
a
clear-cut
segregation
of
long-lived
and
short-lived
groups
in
subsequent
generations.
2.
In
its
genetic
behavior
duration
of
life
is
cora
pletely
and
invariably
associated
with
certain
morpho-
logical
characteristics
of
the
fl
ies,
in
the
sense
that
no
vestigial
-winged
fl
y
has
ever
been
found
to
be
long-
lived,
and
no
group
of
normal
-winged
fl
ies
have
ever
(14)
Vest
d
Wild
9
(44)
F
Wild
Z
d
Wild
Z
9
(46)Wild
r
Vest
03)
been
found
to
have
a
life
curve
even
approaching
in
form
that
which
is
characteristic
of
vestigial
-winged
fl
ies.
(l4)
Vest
Wild
cf
(44)
Wilg1P(1
51
Wild
Pp
(46)Wild
Vest
(l
1_1(a
6
There
are
three
possible
interpretations
of
these
re-
sults
which
suggest
themselves
in
the
light
of
our
pres-
ent
knowledge
of,
and
viewpoint
about,
genetic
phenom-
ena
in
general.
These
are:
1.
It
may
be
assumed
that
duration
of
life
is
deter-
184
TUE
AMERICALV
NATURALIST
[Vor_
LVII
mined
by
a
single
Mendelian
gene,
completely
or
per-
fectly
linked
to
the
gene
for
vestigial,
so
that
crossing
over
never
occurs
between
the
former
and
the
latter.
2.
It
may
be
assumed
that
duration
of
life
is
a
charac-
ter,
like
body
size,
controlled
by
n
factors,
each
of
which
Mondelizes.
This
postulates
what
amounts
to
a
whole
fl
ock
of
genes
for
viability
hi
a
graded
series,
such
that
the
effect
of
each
translates
itself
into
actuality
at
vary-
ing
points
in
the
entire
span
of
imaginal
life.'
(/4)
Vest
d
(52)
Wild
Z
t-
Wild
9
(4d)
(l4)
Vest
g
Wild
d
(44
)
Wild
P
g
(51)
Wild
P
(36)
Wild
O
n
Vest
D,,
(19)
(25)
Wild
Vest
(7)
(31)
Wild
Vest
(
8
)
Me
7
3.
It
may
be
assumed
that
duration
of
life
is
not
in
and
of
itself
a
separate
and
different
character
of
the
organism,
but
instead
is
simply
the
expression
in
time
of
the
organization
of
the
body,
which
organization
may
be
regarded
as
physico-chemical,
structural,
physiologi-
cal
or
functional,
depending
upon
which
of
the
aspects
of
organization
we
may
at
the
time
be
emphasizing.
2
Obviously
if
we
%TO
to
have
viability
or
lethal
genes
at
all
there
is
no
reason
why
we
should
suppose
them
confined
to
embryonic,
or
larval,
or
pupal
stages
of
the
life
history.
A
gene
that
kills
at
age
50
days
is
just
as
easily
conceived
)
and
for
all
we
know
just
as
probable
a
priori,
as
ono
that
kills
2
-
day
-old
lame.
Presumably
what
a
lethal
gene
moans
biologically
is
merely
that
it
determines
an
organizatiou
of
the
body
such
that
viability
ceases
(or
put
the
other
way-
about,
that
the
individual
cannot
go
on
living)
after
that
organization
(structural
or
functional)
has
attained
its
full
somatic
development.
No.
349]
STUDIES
ON
TUE
DURATION
OF
LIFE
185
Of
these
alternative
hypotheses
the
fi
rst
may
be
dis-
missed
at
once
for
two
reasons,
fi
rst
because
it
is
inher-
ently
improbable
biologically,
and
second
because,
as
will
he
demonstrated
in
the
next
following
paper
in
this
series,
the
same
kind
of
evidence
as
that
adduced
in
this
paper
can
be
brought
forward
to
indicate
a
separate
gene
for
duration
of
life
completely
linked
to
the
genes
for
each
one
of
a
whole
series
of
morphological
charac-
ters,
such
as
eye
color
and
the
like.
(44)
d
d
Wild
Vest
9(04)
(44)Wild
9<
Vest
(l4)
F
(5/)
Wdd
Z
g
(52)
Wile
/
Z
P
Wild
(3
T
)Wild
D
lt
Vest
1)
76
(/7)
(32)
Wild
(z7)
tbad
Vest
(/0)
a.
s
(33)Wild
Vest
(5)
Regarding
the
second
hypothesis
it
may
be
said
to
violate
the
logical
principle
Entia
non
suet
multipli-
eanda
prceter
necessitatem.
For
why
should
we
postu-
late
a
series
of
special
genes
for
duration
of
life,
each
necessarily
(because
of
the
observed
facts)
completely
associated
or
li
nked
with
genes
for
morphological
char-
acters.
To
do
so
seems
to
us
to
serve
no
useful
purpose.
With
the
third
hypothesis,
which_
has
been
discussed
by
one
of
us
elsewhere
(52),
the
existing
knowledge
186
THE
AMERICAN
NATURALIST
[Vol-
IATEI
seems
to
he
in
general
in
entire
accord.
It
means
that
in
a
strict
sense
we
can
not
speak
of
the
inheritance
of
duration
of
life
any
more
properly
than
we
can
speak
of
the
inheritance
of
tuberculosis.
3
(l4)
Vest
d
I
Wild
9
(44)
WI
Z
c
52
Wrld
Z
(36)Wild
C
7
,
Vest
C
z
,
(/6)
FIG..
0
What
is
inherited
in
both
eases
is
not,
the
thing
itself,
but
the
bodily
organization,
of
which
duration
of
life,
on
the
one
hand,
and
the
disease
tuberculosis,
on
the
other
hand,
are
in
part
the
manifestations.
in
short
what
is
Vest
8
(14)
Vest
d
Wi/dZ
7
V
11
id
29
VestCss(//)
o.
0
Wild
9
(44)
3
Actually
it
is
convenient
to
speak
of
the
inheritance
of
duration
of
life,
just
as
it
is
to
say
that
tuberculosis
is
in
some
degree
inherited.
And
no
harm
din
be
done
by
such
usage
of
words,
provided
one
understands
that
it
is
merely
a
verbal
economy,
to
save
the
use
of
a
longer
and
more
awkward
phrase.
No.
6491
STUDIES
ON
TEE
DURATION
OF
LIFE
187
inherited
is
the
diathesis
for
longevity,
not
longevity
itself.
Put
in
another
way,
what
this
view
of
the
matter
says
is
that
if
the
duration
of
an
individual's
life
is
an
implicit
function
of
the
individual's
organization
or
COIF
stituticm
(a
purely
physical
thing
fundamentally)
which
(14)
Vest
g
Wild
d
(44)
P
P
9
f;
(2a
Win
4
a3
Vest
A8
3
(10)
no.
11
is
inherited,
it
follows
that
duration
of
li
fe
will
behave
exactly
as
though
it
were
itself
inherited.
Tt
then
di-
rectly
follows
that,
from
the
standpoint
of
methodology,
we
may
by
a
genetic
study
of
duration
of
life
get
valu-
able
knowledge
as
to
the
mode
or
form
of
the
functional
nest
d
('4)
vest
9
-
Wild
cl
(44)
wild
P
w;laPg
(33)Wild
A82
Vest:982
04
relationship
between
physical
organization
and
dura-
tion
of
life,
which
in
fact
is
thus
perceived
from
a
new
angle
to
be
the
fundamental
problem
involving
duration
of
life.
188
THE
A
211.ERIC
AN
NATURALIST
LVII
This
means
that
in
the
present
series
of
-
experiments
we
have
demonstrated
that
the
gene
for
vestigial,
be-
sides
affecting
the
form,
size
and
general
functional
organization
of
the
wings
of
Drosophila,
also
influences,
(14)Vest
d
-Wild
?(44)
Wild
d
z
d
Wild
z
g
(
.
33)
Wild
6
80
FIG.
13
in
and
through
its
effects
on
the
whole
organization
of
every
fl
y
in
which
it
is
present,
the
duration
of
li
fe
of
that
fl
y.
We
have
measured
with
considerable
exact-
ness
and
rnanifoldness
what
the
quantitative
effect
of
(/4)Vest
d
-
Wild
?(44)
t;
Wild
Z
Wild
Z
(34)Wild
Gm
FIG.
14
this
gene
is
upon
duration
of
life
(and,
of
course,
at
the
same
time
the
effect
of
its
allelomorph
the
gene
for
nor-
mal
wing).
This
view
of
the
ease
seems
to
us
to
be
in
accord
with
the
best
current
opinion
as
to
the
biological
meaning
No.
649]
STUDIES
ON
THE
DURATION
OF
LIFE
189
of
mutant
genes
generally.
Thus
Morgan
(9,
pp.
238-
239)
says:
If
we
examine
almost
any
mutant
race,
such
as
the
race
of
white
-eyed
Drosophila,
we
fi
nd
that
the
white
eye
is
only
one
of
tho
characteristics
that
such
a
mutant
race
shows.
The
produc-
tivity
of
the
individual
is
also
much
affected,
and
the
viability
is
lower
than
in
the
wild
fly.
All
of
these
peculiarities
are
found
whenever
the
white
eye
emerges
from
a
cross
and
are
not
separable
from
the
white
-eyed
condition.
It
follows
that
whatever
it
is
in
the
germ
-plasm
that
produces
white
eyes
also
produces
other
modi-
fi
cations
as
well,
and
modifies
not
only
such
"superficial"
things
as
color,
but
also
such
"fundamental"
things
as
productivity
and
viability.
Many
examples
of
this
manifold
effect
are
known
to
students
of
heredity.
It
is
perhaps
not
going
too
far
to
say
that
any
change
in
the
germ
-plasm
may
produce
many
kinds
of
effects
on
the
body.
Clearly
then
the
character
that
we
choose
to
follow
in
any
case
is
only
the
most
conspicuous
or
(for
purposes
of
identification)
the
most
strik-
ing
or
convenient
modification
that
is
produced.
Since,
however,
these
effects
always
go
together,
and
can
he
(explained
by
the
as-
sumption
of
a
single
unit
difference
in
the
germ
-plasm,
the
partic-
ular
difference
in
the
germ
-plasm
is
more
significant
than
the
character
chosen
as
its
index.
The
general
view
as
to
the
mechanism
through
which
duration
of
life
behaves
as
an
inherited
character
which
is
here
tentatively
adopted
has
been
discussed
in
more
p4)
Vest
Wild
d
(44;
fyi
Wad
(
P
d
FIG.
35
Wild
Pp
67,1Wild
E
s
.;
detail
by
Pearl at
various
points
in
"The
Biology
of
Death"
(52).
Thus
it is there
suggested
(p.
21.2)
that
"what
heredity
does
in
relation
to
duration
of
life
is
chiefly
to
determine,
within
fairly narrow
limits,
the
total
190
THE
AMERICAN
NATURALIST
[Voi,.
LVII
energy
output
which
the
individual
can
exhibit
in
its
life
time.
This
limitation
is
directly
brought
about
pre-
sumably
through
two
general
factors:
viz.,
(a)
the
kind
or
quality
of
material
of
which
this
particular
vital
ma-
chine
is
built,
and
(b)
the
manlier
in
which
the
parts
are
put
together
or
assembled."
(/4)Vest
Wild
c1(44)Wild
F
Wild
P
(so)
W
E85
Ma.
13
P
In
this
statement
(a)
and
(b)
are
equivalent
to
what
is
here,
in
more
general
terms,
called
"organization."
To
come
110w
to
details
of
the
present
investigation
there
are
a
number
of
points
for
which
we
have
no
pres-
ent
explanation
to
offer.
The
chief
of
these
are:
1.
The
significantly
aberrant
Mendelian
ratios
of
vestigial.
2.
The
peculiar
differences
in
duration
of
life,
ap-
parently
associated
with
these
departures
from
Mende-
lian
expectation
of
vestigial.
We
refer
to
such
matters
as
the
divergent
results
of
Z
X
P
as
compared
with
other
inter
se
matings
of
I',
to
produce
F
2
;
the
peculiar
behavior
of
duration
of
life
in
back
crosses:
the
failure
of
reciprocal
makings
to
give
identical
results,
though
all
morphological
mutants
dealt
with
have
their
location
in
the
second
chromosome.
3.
The
steady
degradation
in
mean
duration
of
life
after
F,.
The
present
material
is
inadequate
to
the
solution
of
th.ese
difficulties.
It
has
certain
defects
which
contribute
No.
6491
STUDIES
ON
THE
DURATION
OF
LIFE
191
heavily
to
this
inadequacy.
In
the
first
place
Quintuple
carries
5
different
mutations
each
one
of
which
has
some
influence
on
duration
of
life,
as
will
be
demon-
strated
in
the
next
following
paper
in
this
series.
We
took
account
of
but
one
of
these
morphological
charac-
ters
(namely
vestigial)
in
this
study.
In
the
second
place
the
material
is
not
large
enough
in
mass
to
deal
adequately
with
the
kind
of
deviations
from
eqoecta-
lion
that
appeared.
In
subsequent
investigations
we
hope
to
remedy
these
defects.
In
the
meantime
the
present
study
has
fulfilled,
we
believe,
the
highest
prag-
matic
test
of
any
genetic
inquiry.
Namely,
we
are
now
in
a
position
to
predict
what
the
duration
of
life
of
a
progeny
group
of
Drosophila
will
he,
from
a
knowledge
of
the
duration
of
life
of
the
parents.
SummARy
In
this
study,
involving
the
determination
of
the
dura-
tion
of
life
of
5,415
individual
fl
ies,
a
cross
was
made
between
a
long-lived
stock
of
Drosophila
(Old
Falmouth,
wild
type)
and
a
short-lived
stock
(Quintuple).
In
the
P,
generations
the
progeny
were
somewhat
longer
lived
than
either
parent
stock.
In
F,
there
was
a
definite
and
clear-cut
segregation
in
respect
of
duration
of
life,
long-
lived
and
short-lived
groups
reappearing,
with
virtually
identical
mean
duration
of
life
to
those
of
the
original
parent
stocks.
In
its
genetic
behavior
duration
of
li
fe
was
found
to
be
completely
and
invariably
associated
with
certain
morphological
characteristics
of
the
organ
-
islet,
in
the
sense
that
no
vestigial
-winged
fl
y
has
ever
been
found,
in
the
entire
experience
of
the
laboratory,
to
be
long-lived,
and
no
group
of
normal
-winged
fl
ies
has
ever
been
found
to
have
a
life
curve
even
approach-
ing
in
form
that
which
is
characteristic
of
vestigial
-
winged
flies.
The
probable
meaning
of
these
results,
in
relation
to
genetic
phenomena
in
general,
is
discussed.
192
THE
AMERICAN
NATURALIST
[Vol..
LVII
LITERATURE
CITED
(The
plan
of
numbering
citations
is
explained
in
the
second
of
these
Studies.
AMER.
NAT.,
Vol.
56,
p.
174.)
52.
Pearl,
R.
The
Biology
of
Death.
Philadelphia
(.J.
B.
Lippincott
Co.),
1922.
53.
Id.
A
Note
on
the
Inheritance
of
Duration
of
Life
in
Man.
Amer.
Jour.
Hug.,
Vol.
2,
pp.
229-233,
1922.
54.
Id.
A
System
of
Recording
Types
of
Mating
in
.Experimental
Breeding
Operations.
Salome,
N.
S.,
Vol.
42,
pp.
383-386,
1916.
.55.
East,
E.
M.,
and
Tones,
D.
F.
Inbreeding
and
Outbreeding,
their
Genetic
and
Sociological
Significance.
Philadelphia,
1919.
56.
Shull,
G.
II.
The
Genotypes
of
Maize.
ATER.
NAT.,
Vol.
45,
pp.
234-
252,
1911.